|
|
small (250x250 max)
medium (500x500 max)
Large
Extra Large
Full Size
Full Resolution
|
|
CENTER FOR NUTRITION POLICY AND PROMOTION Ann M. V eneman, Secretary U.S. Department of Agriculture Eric M. Bost, Under Secretary Food, Nutrition, and Consumer Services Steven N. Christensen, Acting Deputy Director Center for Nutrition Policy and Promotion P. Peter Basi otis, Director Nutrition Policy and Analysis Staff The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, gender, religion, age, disability, political beliefs, sexual orientation, or marital or family status. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communication of program information (Braille, large print, audiotape, etc.) should contact USDA's TARGET Center at (202) 720-2600 (voice and TDD). To file a complaint of discrimination, write USDA, Director, Office of Civil Rights, Room 326-W, Whitten Building, 14th and Independence Avenue, SW, Washington, DC 20250- 9410 or call (202) 720-5964 (voice and TDD). USDA is an equal opportunity provider and employer. Editor Julia M. Dinkins Assistant Editor David M. Herring Features Editor Mark Lino Managing Editor Jane W. Fleming Contributor Joan C. Courtless Family Economics and Nutrition Review is writlen and published each quarter by the Center far Nutrition Policy and Promotion, U.S. Department of Agriculture, Washington, DC. The Secretary of Agriculture has determined that publication of this periodical is necessary in the transaction of the public business required by law of the Department. This publication is nat copyrighted . Contents may be reprinted without permission, but credit to Family Economics and Nutrition Review would be appreciated. Use of commercial or trade names does nat imply approval or constitute endorsement by USDA Family Economics and Nutrition Review is indexed in the following databases: AGRICOLA, Ageline, Economic Literature Index, ERIC. Family Studies, PAIS, and Sociological Abstracts. Family Economics and Nutrition Review is for sole by the Superintendent of Documents. Subscription price is $19.00 per year ($23.75 for foreign addresses). Send subscription order and change of address to Superintendent of Documents, P.O. Box 371954, Pitlsburgh, PA 15250-7954. (See subscription form on p. 127.) Original manuscripts are accepted far publication. (See "guidelines for authors" an back inside cover.) Suggestions or comments concerning this publication should be addressed to Julia M. Dinkins, Editor, Family Economics and Nutrition Review, Center far Nutrition Policy and Promotion, USDA. 31 01 Park Center Drive, 1Oth Floor, Alexandria, VA 22302. The Family Economics and Nutrition Review is now available at http:/ /www.cnpp.usda.gov. (See p. 126.) Research Articles 3 Profiles of Selected Target Audiences: Promoting the Dietary Guidelines for Americans Kay Loughrey, P. Peter Basiotis, Claire Zizza, and Julia M Dinkins 15 U.S. Teens and the Nutrient Contribution and Differences of Their Selected Meal Patterns Anna Maria Siega-Riz, Claude Cavadini, and Barry M. Popkin 27 Associations Between the Milk Mothers Drink and the Milk Consumed by Their School-Aged Children Rachel K. Johnson, Celeste V. Panely, and Min Qi Wang 37 The Effects of Food Advertising Policy on Televised Nutrient Content Claims and Health Claims Carol Byrd-Bredbenner and Darlene Grasso 50 The Thrifty Food Plan, 1999: Revisions of the Market Baskets Staff at the Center for Nutrition Policy and Promotion-Compiled by Mark Lino 65 Sample Menus and Recipes Based on the 1999 Thrifty Food Plan Staff at the Center for Nutrition Policy and Promotion- Compiled by Myrtle Hogbin and Mark Lino Research Briefs 77 USDA's Food Guide: Updating the Research Base to Reflect Changes in Food Consumption Patterns Kristin L. Marcoe 81 USDA's Expenditures on Children by Families Project: Uses and Changes Over Time Mark Lino 87 Current Knowledge of the Health Effects of Sugar Intake Anne L. Mardis 92 Insight 11 : Food Portions and Servings: How Do They Differ? Myrtle Hogbin, Anne Shaw, and Rajen S. Anand 95 Insight 14: A Focus on Nutrition for the Elderly: It's Time to Take a Closer Look Nancy W. Gaston, Anne Mardis, Shirley Gerrior, Nadine Sahyoun, and Rajen S. Anand 98 Insight 19: Beliefs and Attitudes of Americans Toward Their Diet Julia M. Dinkins Volume 13, Number 1 2001 Research Summaries 101 Consumer Price Index Research Series Using Current Methods 1 05 Changes in the Health Services Industry 108 Extended Measures of Well-Being: Meeting Basic Needs 112 Measuring Time at Work Regular Items 1 15 Research and Evaluation Activities in USDA 118 Federal Studies: Review of the Nutritional Status of WIC Participants 120 Journal Abstracts 122 Official USDA Food Plans: Cost of Food at Home at Four Levels, U.S. Average, May 2001 123 Official USDA Alaska and Hawaii Thrifty Food Plans: Cost of Food at Home (2nd Half 2000) 124 Consumer Prices 125 U.S. Poverty Thresholds and Related Statistics Kay Loughrey, MPH, RD MRP P. Peter Basiotis, PhD Center for Nutrition Policy and Promotion Claire Zizzo, MS University of North Carolina at Chapel Hill Julia M. Dinkins, PhD Center for Nutrition Policy and Promotion 2001 Vol. 13 No. I Research Articles Profiles of Selected Target Audiences: Promoting the Dietary Guidelines for Americans To decrease the risk of nutrition-related diseases, Americans need to narrow the gap between scientifically based nutrition guidance and their nutrition-related behaviors. This study examines the usefulness of segmentation and audience-profiling techniques in promoting the Dietary Guidelines, designed to help narrow this gap. Using the 1991- 94 survey of the Market Research Corporation of America Information Services (MRCA), we segmented 491 women gatekeepers into tertiles (Better Eaters, Fair Eaters, and Poor Eaters) based on their scores on a modified version of the Healthy Eating Index. We then compared the segments' demographic characteristics; health and diet orientation; values about, and perceived benefits and barriers to, healthful eating; nutrition, food preparation, and shopping habits; and media habits. Results showed that women gatekeepers were interested in improving their diets, and they differed significantly regarding values, benefits, and barriers of eating a healthful diet and nutrition; food preparation practices; and shopping habits. We discussed the implications of these differences in terms of improving the quality of the diet. T he Dietary Guidelines for Americans (23), issued by the U.S. Departments of Agriculture (USDA) and Health and Human Services (DHHS), answer this basic question: "What should people eat to stay healthy?" Forming the basis of Federal nutrition policy affecting food, nutrition education, and information programs, the Guidelines stress the significance of dietary balance, variety, and moderation (7). Still, in the United States, four of the leading causes of death-heart disease, cancer, stroke, and diabetes-are linked to nutrition (1 0). Americans still need to increase total intake of fruits, vegetables, and grain products and to decrease intake of fat and saturated fat. Although some progress has been made based on progress in meeting Year 2000 Objectives, the startling increase in the portion of Americans who are overweight or obese poses one of the biggest challenges in meeting Healthy People 2000 (24). A summary measure of dietary status- the Healthy Eating Index-has shown that 7 of I 0 Americans need to improve their diet (4) . Other results have also indicated that although Americans choose a wide variety of foods, they consume less than the recommended servings from the fruit, dairy, meat, grains, and vegetable groups of the Food Guide Pyramid. Americans' consumption of calories from fats and sugars, however, exceeds Pyramid recommendations (11). 3 Thus evidence has shown that Americans still need to improve their diet; Americans need to narrow the gap between scientifically based nutrition guidance and consumer behavior that may increase the risk of illness from nutrition-related diseases. To better meet the needs of the public, some authors believe the Guidelines need to do two things: (1) continue to advance national dietary guidance that is based upon scientific evidence and (2) promote dietary guidance in ways that will lead to behavior change, improved health, and nutritional well-being (22) . The purpose of this study is to examine the extent to which major differences exist between audience segments on key variables, to profile these audience segments, and to suggest whether these differences warrant distinct nutrition education approaches in attempting to change dietary behaviors. We describe three segments of female gatekeepers and how their characteristics differ on several dimensions: demographic and health status; values about, and benefits and barriers to, healthful eating; nutrition, food preparation, and shopping habits; and media habits. We discuss the implications of these differences in terms of improving the dietary behavior of these segments. We believe that nutrition educators can directly apply this information when they design program interventions. The underlying assumption of social marketing and marketing approaches is that different audience segments require alternate approaches for achieving a desired behavior change. This study examines whether this assumption applies to nutrition education to create dietary behavior change. 4 Lastly, we examined results in relationship to behavioral models and theories. We examined how the segments might differ with respect to their stage of behavior change and the extent to which audience segments could be described, based on Prochaska and Di Clemente's trans theoretical model of change (18). The stages of this model are precontemplation (not considering whether to make a change), contemplation (thinking about making a change), decision (making definite plans to change), action (initiating change), and maintenance. This model has been used to describe dietary behavior in relationship to weight control and the reduction of dietary fat (6, 19). We looked to social learning theory, which is based on social cognitive theory, to inform recommendations for designing strategies for behavior change (3,17). Social learning theory emphasizes the interaction of cognition, other personal factors (e.g., selfefficacy), and environmental factors on behavior. Several critical personal factors suggested by social learning theory have been assessed in this analysis: • Perception of the situation • Anticipated outcomes of behavior • Knowledge and skills to perform a behavior • Confidence in performing a behavior We considered the theory of planned behavior in forming program implications (1). This theory suggests that people will be more likely to take action if it leads to consequences they desire. It also suggests that behavior and behavioral intent are influenced by the degree of control people think they have over circumstances and their ability to perform a behavior. Background Research indicates that nutrition promotion of the Guidelines should focus on behavior change; have a strong consumer orientation; segment and target consumers; use multiple, reinforcing, interactive channels; and refine consumer messages continually (22,23). Segmentation, a frequently used approach in commercial-sector marketing, has been used in programs designed to change health behaviors (2) and has been used to create a profile or snapshot that represents the target audience. It, as well, has encouraged creative communication that is tailored to the target audience (6,12,15). To segment audiences, social marketers analyze potential markets and create subgroups of target populations with similar characteristics regarding the desired behavior. Then they allocate resources among one or more subgroups and vary the methods used to reach each subset (2). Health communicators also use segmentation methods to identify people who are similar in key respects and to tailor the content and delivery of the communication based on people's profiles (16,21). Target-audience profiles have been used in large-scale nutrition education programs, including the 5 A Day media campaign of the National Cancer Institute (1 3,15) and the Nutrition and Physical Activity program of the Centers for Disease Control (9). Family Economics and Nutrition Review Methods Database We analyzed data from the 1991-94 survey of the Market Research Corporation of America Information Services (MRCA). Nationally representative, the MRCA survey consists of information on people's food and beverage consumption and their opinions and attitudes about general interests, health, diet and food preparation, shopping, and media usage. The MRCA data set consists of five surveys and two database systems: Household Information Form, Menu Census Diaries, Pyschographic Questionnaire, Diet Information Quiz, and Food and Nutrition Attitude Inventory. To select participating households, MRCA uses a multistage, stratifiedrandom procedure. In stage !-the Household Recruiting Pool-a sampling pool of households is generated from generic consumer listings of U.S. households of various demographic types. Households that agree to participate then qualify for the second stage of sampling-the National Consumer Panel. The Panel consists of 5,000 households whose demographic characteristics (household size, homemaker age, household income, census regions, and metro-area size) are matched to the U.S. Census. The third stage-the Menu Census Panelconsists of a subsample of households (n=2,000) from the National Consumer Panel. For the Menu Census Panel, MRCA uses a stratified-random procedure to select 500 households each quarter. Detailed food diaries of food and beverage consumption are collected for 14 consecutive days. Actual serving sizes are not collected. They are imputed based on eating occasions for individual foods by applying standard serving sizes. For this reason, they should be considered 2001 Vol. 13 No. 1 estimates rather than precise measures of food and beverage consumption. The Nutrient Intake database measures macro- and micro-nutrient intake; the Food Guide Pyramid database measures "servings"' of the Pyramid Food Groups. Healthful Eating Measure The USDA Healthy Eating Index (HEI) measures the overall quality of Americans' diet (4) and uses data from the USDA Continuing Survey ofFood Intakes by Individuals (CSFII). The HEI uses 1 0 components to measure different aspects of a healthful diet: • Components 1-5 measure the degree to which a person's diet conforms to serving recommendations of the food groups of the USDA Food Guide Pyramid: Grains (bread, cereal, rice, and pasta), vegetables, fruits, milk (milk, yogurt, and cheese), and meat (meat, poultry, fish, dry beans, eggs, and nuts). • Components 6 and 7 measure consumption of total fat and saturated fat, respectively, as a percentage of total food energy intake. • Component 8 measures total cholesterol intake. • Component 9 measures sodium intake. • Component 10 measures the variety of a person's diet on any given day. 1 MRCA used total frequency of"eatings" as the main measure of the individual food consumed. MRCA estimated serving sizes for each eating occasion for over 330 collapsed food categories based on 1987-88 USDA data on number of grams for each eating occasion for individual food items. MRCA then assigned different serving sizes to 18 age-gender groups: four age groups for children under 12 and seven age groups each for males and females over age 13. Americans still need to improve their diet; Americans need to narrow the gap between scientifically based nutrition guidance and consumer behavior that may increase the risk of illness from nutrition-related diseases. 5 Each component of the HEI has a maximum score of 10 and a minimum score of zero; intermediate scores are computed proportionately. The maximum overall score for the 10 components combined is 100. Higher component scores indicate intakes close to recommended ranges or amounts. The MRCA does not provide information on variety; hence, we used a modified version of the HEI to examine characteristics that distinguish women from the MRCA sample with higher quality diets from those with lower quality diets. All scores on the modified version were adjusted to a 1 00-point score. Thus the total maximum score was 100. To compute individual HEI scores, we matched the female gatekeeper to the appropriate serving recommendations of the Pyramid Food Groups. We calculated gatekeepers' average percentage of calories from total fat and saturated fat and compared their intakes of cholesterol and sodium with Pyramid recommendations. Sample We selected healthy adult women in the United States as the unit of analysis (target audience) because they often are gatekeepers who shape their family's nutrition and health habits. Our sample consisted of women gatekeepers aged 25 through 55, reporting household income of $20,000 to $125,000 and no major health problems. Those excluded reported having high blood pressure, diabetes, heart disease, high levels of serum cholesterol, or followed a diet for diabetes or allergies. We could not use marital status as a screening variable because MRCA does not include information on respondents' marital status. The database also does not include information on vegetarian diets, employment status or profession, 6 and the relationship of household members. The final sample of 491 gatekeepers was weighted to reflect the U.S. population of interest. After ranking and dividing the gatekeepers into tertiles (segments) based on their scores on the modified HEI, we developed profiles of the women gatekeepers and used multiple t tests to examine differences among the three segments. SUDAAN (Software for the Statistical Analysis of Correlational Data), which accounts for sampling designs that are complex and stratified, was used in the analysis to ensure appropriate estimates of standard errors for hypotheses testing.2 Results Demographic Characteristics The women gatekeepers who were Better Eaters (having the highest HEI score) are the basis of comparison with other groups of women gatekeepers: Fair Eaters and Poor Eaters. The women gatekeepers differed in some ways (table 1). Compared with the other groups, the Better Eaters more closely met the recommendations of the USDA Food Guide Pyramid. Based on percentages, overall, the women gatekeepers' average Healthy Eating Index score was 57 percent. With an average score of74 percent, the Better Eater had the higher HEI score, followed by the Fair Eater, with 62 percent; and Poor Eater, with 52 percent. Healthy Eating Index scores were calculated based on the degree to which a person in the sample's diet 2 "SUDAAN is specifically designed for analysis of cluster-correlated data from studies involving recurrent events, longitudinal data, repeated measures, multivariate outcomes, multistage sample designs, stratified designs, unequally weighted data, and without replacement samples" (20). conformed to serving recommendations of the food groups of the USDA Food Guide Pyramid as previously described. There are small differences in the gatekeepers' average years of education, height, Body Mass Index (BMI), likelihood of having children present in the household, and race. The Better Eater was more likely than the other Eaters to have more years of education. Compared with the Poor Eater, the Better Eater had a lower BMI score, was slightly taller, and more likely to be White or of a race other than Black. Compared with the Fair Eater, the Better Eater was less likely to have children. The women gatekeepers had some characteristics in common (tables 1 and 2). Their characteristics were considered similar if more than 60 percent of the women in each group exhibited them and if the differences in the characteristics were statistically insignificant (p>.Ol). These three groups were similar demographically based on age, household size, household income, and self-reported weight. Values, Benefits, and Barriers to Healthful Eating Similar to the Better Eater, the Fair Eater (F) reported that eating a healthful diet was important to her (table 3). Both said they could avoid future health problems-a perceived longterm benefit-by eating more healthfully. Similarly, the Fair Eater and the Better Eater reported that eating "healthy foods" gave them the energy they needed- a perceived shortterm benefit- and agreed that eating "healthy foods" improved their physical appearance. Family Economics and Nutrition Review Table 1 . Education distinguishes all three segments of women gatekeepers: Demographic and health status variables, MRCA 1991-94 Diet status Variable Better Eaters Fair Eaters Poor Eaters Mean Age (years) 39 38 38 Household size 3.3 3.3 3.3 Household income (thousands) 42.97 41.40 41.93 Education (years) 14.2* 13.7* 13.2* Weight (kg) 67.39 67.96 71 .65 Height (em) 164.7* 163.9 162.7* BMI 25.07* 25.54 27.31* Percent HEI score' 74 62 52 Children present 56* 72* 65 White 94.9* 87.5 83.6* Block 4.1 7.3 9.3 Other 1.0* 5.2 7.1* 'The Healthy Eating Index scores differ because this foetor was used to segment the women gatekeepers. *Means or percentages within the some row ore significantly different (p < 0.05). The Fair Eater differed, however, from the Better Eater in two important ways. ( 1) She was Jess likely than the Better Eater to believe she could avoid future health problems by exercising. (2) Both convenience and taste were barriers for the Fair Eater, who was more likely than the Better Eater to say that "healthy foods" had to be convenient for her to use them and to report that a reason for not choosing healthful foods was because they didn't taste good. The Poor Eater (P) was less likely than the Better Eater to believe it was important to eat a healthful diet, look and feel physically fit, maintain a proper weight, and to identify with potential benefits of healthful eating. She was less likely to agree that she could avoid future health problems by eating a healthful diet and by exercising; she was less likely to report the perceived short-term benefit that eating "healthy foods" gave her the energy she needed and improved her physical appearance. The Poor Eater also 2001 Vol. 13 No. 1 indicated that she was less likely than her counterpart to say she knew how to eat healthfully. She was, however, more likely than the Better Eater to report that eating healthfully was too complicated and confusing. Health and Diet Orientation All of the women gatekeepers believed they were knowledgeable about health and nutrition (table 2). They reported an interest in improving their diets, agreed they had some weight to lose, and tried to do so, at least occasionally. Similarly, they agreed that it was important for them to live long and healthy lives. Nutrition, Food Preparation, and ·Shopping Habits Similar practices among the women gatekeepers extended to how they shopped for food and planned and prepared it (table 2). Among the many similarities, all three groups redeemed the coupons they clipped from magazines and newspapers. The Poor Eater was less likely than the Better Eater to believe it was important to eat a healthful diet, look and feel physically fit, maintain a proper weight, and to identify with potential benefits of healthful eating. 7 Table 2. Better Eaters, Fair Eaters, and Poor Eaters have many characteristics' in common, MRCA 1991-94 Variable Health and diet orientation Physical activity Pyschographics Shopping Food planning and preparation Family eating habits Media Commonalities Believe they ore knowledgeable about health and nutrition Interested in improving their diets Think they hove some weight to lose Try, at least occasionally, to lose weight Believe it is important for them to live a long, healthy life Frequency Like to meet new people Join actively in community groups Desire to be well respected Like the outdoors Enjoy taking the family to a different vocation spot each year Make a complete list before going shopping Enjoy browsing through supermarket aisles Do not like the excitement of a busy supermarket Save a lot of money by shopping around for food bargains Stock up on named brand foods that they like during soles Cut coupons out of newspapers and magazines Redeem coupons (almost always) Send away for items offered through advertising Willing to pay for certain food items for special occasions En joy cooking and think of themselves OS creative cooks Don't like to bother cooking just for themselves (when alone) Enjoy preparing a fancy meal for their families once in awhile Collect recipes from the food sections of the newspapers Exchange recipes with friends and relatives Add something extra (almost always) to prepared foods Serve the some evening meals from one week to the next Try to make use of leftovers but usually throw them out Hove some family members who ore concerned about being overweight View television-network evening news, cable news/ television Read magazines and newspaper 'Characteristics were common if more than 60 percent of each group exhibited them and if the differences in the characteristics were statistically insignificant (p > .01 ). 8 Family Economics and Nutrition Review Table 3. Most measured beliefs and practices of Poor and Fair Eaters differ from those of Better Eaters, MRCA 1991-94 Degree to which Poor (P) and Fair (F) Eaters soy the following, compared with Better Eaters Variable As likely More likely Less likely Values, Benefits, and Batriers Eating a healthy diet is important to me. F I can ovoid future health problems by eating healthfully. F I choose healthy foods because they give me the energy I need. F I choose healthy foods because they improve my physical appearance. F Healthy foods hove to be convenient for me to use them . A reason for not choosing healthy foods is they don't taste good. Trying to eat healthy is too complicated and confusing . I con ovoid future health problems by exercising. It is important for me to look and feel physically fit. It is important for me to maintain my proper weight. I know how to eat healthy. Nutrition, Food Preparation, and Shopping Habits I worry about the nutritional content of the foods I eat. F I always see to it that my family takes vitamins. P I'm much more willing to try a new recipe when someone I know tried it and liked it. I always or usually pay attention to on-shelf, aisle display. Most snack foods I like ore unhealthy. I do not discuss various foods and their food values with my family so they understand nutrition better. I always pay attention to instant coupons. I make every possible effort to see that my family eats really nourishing foods. I get upset if the family doesn't eat together. I go out of my way to buy non-fat foods. Frozen foods ore more nutritious than conned foods. I serve fish because it has less fat. I disagree that red meat is better for your health than fish . I do not look for prepared dishes when I shop. I collect recipes from magazines. I disagree that my family is easy to please. Media I watch television in general, including entertainment programs, and daytime television . I watch television programs like police/ private eye and daytime serials because I really like them . I watch television serials/soap operas because I like them . I watch prime-time television programs. I read women's general interest magazines. F F F, p F F p p P, F F, p F, p F P, F p p p F, p p p p p F F, p F, p F, p F, p F, p F p p p p Note: The "F" and "P" for the Fair Eaters and Poor Eaters, respectively, indicate that these women gatekeepers differ significantly from the comparison group: the Better Eaters, at the 0.01 level. 2001 Vol. 13 No. 1 9 Compared with the Better Eater, the Poor Eater was less likely to worry ~bout the nutritional content of the foods she ate. 10 The groups differed, however, in a number of important ways related to nutrition, food preparation, and shopping habits (table 3). Similar to the Better Eater, the Fair Eater worried about the nutritional content of the foods she ate. Still, she was less likely than the Better Eater to make an effort to serve her family nourishing foods, get upset if the family didn't eat together, and go out of her way to buy nonfat foods. She was more likely than the Better Eater to pay attention to on-shelf, aisle display ads and instant coupons and to look for prepared foods when shopping. Compared with the Better Eater, the Poor Eater was less likely to worry about the nutritional content of the foods she ate. Like the Fair Eater, she was also less likely than the Better Eater to make every possible effort to see that her family ate nourishing foods, to get upset if the family didn't eat together, and to go out of her way to buy nonfat foods. The Poor Eater was more likely than the Better Eater to pay attention to instant coupons, to agree that most of the snack foods she liked were unhealthful, and to disagree that she discussed foods with her family so they understood nutrition better. Media The three groups watched similar television programs or stationsevening network news, cable news, and cable TV-and they read similar magazines and newspapers (table 2). However, the Fair Eater and Poqr Eater were more likely than the Better Eater ~o watch television in general, includmg entertainment (non-news) shows and daytime programs (table 3). The Poor Eater also watched less primetime television than did the Better Eater and was less likely to read women's general interest magazines. Discussion Profiles Demographic differences in audience segments do not explain the overall differences in the three segments' approaches to food consumption. Results of this analysis indicate a small number of demographic differences. Then what might explain these differences? The Better Eaters are more likely than the Poor Eaters to report that eating a healthful diet is important to them and are concerned about the nutritional content of their diets. They are likely to perceive short- and long-term benefits of eating healthfully, and are taking action to eat healthfully. Better Eaters are categorized in this analysis as being either in the action or maintenance stages of the transtheoretical model of change, though direct assessment of the stages of change was not measured in this analysis. Better Eaters are considered in one of these two stages of change based on their HEI score, their concerns about nutrition, and their greater tendency to act on their concerns. It is therefore not possible to determine precisely whether they are in the action or maintenance stage, using the algorithm applied by Curry et al. for staging dietary fat reduction (6). In terms of social learning theory, Better Eaters appear to be able to anticipate the outcomes of their behavior and self-determine their behavior, successfully although not perfectly. They appear to be confident of their ability to carry out healthful eating behaviors based on their being less likely to report that trying to eat more healthfully is complicated and confusing than did women in the other two segments. Better Eaters experience Family Economics and Nutrition Review a rather high degree of control over their circumstances in terms of eating healthfully, based on their responses to all questions, collectively. This characteristic is a key factor in the theory of planned behavior. Still, Better Eaters have room for improving their diets based on their HEI scores. Fair Eaters, compared with Better Eaters, report a mixture of benefits, barriers, and actions that may account for their lower HEI score. Like the Better Eaters, Fair Eaters are more likely than the Poor Eaters to report that eating a healthful diet is important to them, and are concerned about the nutritional content of their diets. They are as likely as Better Eaters to perceive short- and long-tenn benefits of eating healthfully, and are taking some action to eat healthfully. However, they are less likely to go out of their way to eat healthfully, such as making an effort to serve their families nourishing foods and buying nonfat foods. They are more likely to respond to in-store promotions such as on-shelf, aisle display ads, and instant coupons. Taste and convenience are especially important to Fair Eaters, and they are more likely than Better Eaters to select prepared foods. In terms of media use, they are more likely to watch television, particularly for entertainment. Lastly, the Fair Eaters are more likely to report that eating healthfully is complicated and confusing, compared with Better Eaters. In sum: Fair-Eaters are convinced yet not committed to eating healthfully. While they are interested in the positive results associated with eating healthfully and are convinced of its benefits, Fair Eaters are less proactive in making healthful eating choices, and appear to respond passively to stimuli in their environment, be it family, in-store cues, desire for sensory satisfaction, or ease in meal preparation. As a group, they 2001 Vol. 13 No. I appear to eat healthfully when it's convenient and could be characterized as "convinced, but not committed" to eating healthfully. Many factors can intervene in their environment to prevent them from eating healthfully. Fair Eaters could be considered to be in a late stage of contemplation in terms of stages of change, although screening questions for staging were not included in the original MRCA questionnaire. No questions were asked that could help determine whether Fair Eaters had developed a plan of action that would place them in the preparation stage of the transtheoretical model of change. Still, their passivity in relationship to environmental cues indicates that they have not developed a concerted plan of action that they intend to implement in the near future. In terms of social learning theory, Fair Eaters are aware of the outcomes of behaviors, including expected results and benefits but lack the knowledge and confidence to eat more healthfully based on the fact that, compared with Better Eaters, they are more likely to report that trying to eat healthfully is too complicated and confusing. They also seem to experience a rather low degree of control over their circumstances, an important factor influencing their behavior that is emphasized by the theory of planned behavior. A number of factors may prevent Poor Eaters from taking actions that could improve their dietary habits, factors that may account for their HEI scores being the lowest among these three groups. They are less likely to report an interest in achieving results related to healthful eating. For example, they are less likely to report that eating more healthfully is important to them, compared with Better Eaters. Poor Eaters are also less likely to be convinced of long-term benefits: they are less likely than Better Eaters to agree that they can avoid future health problems by eating a healthful diet. Nor are they convinced of short-term benefits such as being less likely to agree that "healthy foods" give them the energy they need. They are also less likely to know how to eat healthfully and are more likely to perceive that eating healthfully is complicated and confusing. Poor Eaters are less concerned about nutrition for themselves and their families: they are less likely to report that they worry about the nutrient content of the food they eat. They are also less likely to talk with their families about foods in terms of their nutritional value or to report making every possible effort to see that their families eat nourishing food. Thus Poor Eaters are somewhat interested in improving their diets, but are not convinced of the benefits of doing so. They are also less concerned with achieving the potential results of eating healthfully than are Better Eaters. While they, like other gatekeepers, claim to be knowledgeable about health and nutrition, they admit to not knowing how to eat healthfully. They could be characterized as "interested but unconvinced" that healthful eating is particularly relevant to them. Poor Eaters could be categorized as being in an early phase of contemplation (transtheoretical model of change) based on their interest in improving their diet. Although Poor Eaters appear to be aware of where they stand when it comes to eating healthfully, they lack three key critical personal factors described by social learning theory: (1) the ability to anticipate outcomes of their behavior, (2) knowledge and skills to act, and (3) confidence to perform this behavior. 11 Program Implications Given the large number of characteristics these three segments of women have in common, should the same approach to nutrition education be used for these three groups? Speaking in favor of a common approach are the characteristics the three segments share. However, many of the characteristics the three audience segments have in common may be attributed to the fact that the segments are all primary food preparers. A number of important differences among these three segments of women discussed in this paper suggest that different approaches to nutrition education are needed for each segment. For the Better Eaters especially, providing tips that are simple, positive, and easy to apply may build on their current interest and actions to improve their diets. A different approach should be used with Fair Eaters. Nutrition education for this group should appeal to their interest in taste and convenience. Communication and education strategies should be used to deliver actionable messages and illustrate easy methods for improving their diet that do not sacrifice taste. Suggestions should be offered that are easy to apply such as adding a grated carrot to prepared tomato sauce as a way to add sweetness, improve its taste, and add important nutrients. It may also be helpful to highlight convenient ways to more healthful eating such as offering ideas that they can do quickly such as a "10-minutea- day" way to improving their eating habits. Fair Eaters should be targeted with a few carefully selected nutrition messages that are easy to understand and apply, and that are likely to cut through confusion generated by media coverage of nutrition news. Nutrition education for Fair Eaters should use mass media to remind them frequently about eating healthfully. It should also be presented in an entertaining way, 12 because this audience is used to regular television entertainment. It will require a highly targeted approach to reach Poor Eaters with nutrition education. An approach is needed that immediately captures their attention and establishes cultural and lifestyle relevance. To help establish relevance of consequences of healthful eating, messages to this audience should come from people they perceive as peers or from someone they admire, such as a celebrity, who can model the desired behavior. For example, the Milk Mustache Campaign has shown celebrities and opinion leaders with their milk mustaches as a way to establish that drinking milk is a highly acceptable and desirable behavior with their target market. Nutrition education programs and materials that are highly targeted to a specific lifestyle or cultural experience are likely to be welcomed. For example, the National Cancer Institute developed and tested Down Home Healthy, a recipe booklet designed for an African-American audience, and found that respondents were highly interested in this book because of its cultural relevance (8). Introducing this recipe booklet was used to explore interest in an approach of encouraging African Americans to use modified versions of traditional recipes to lower fat and increase fiber intake. Responses to the recipe booklet and accompanying brochure were the most active and engaging aspects of focus group sessions. Participants welcome9 this approach, if the taste of the food presented in the recipes met their expectations. Successful nutrition education strategies are recommended that will break abstract nutrition concepts into practical action steps that can easily be mastered and applied to help build knowledge, skills, and self-efficacy for eating more healthfully. For example, guidance about adding more fiber to the diet should include a brief discussion of the Nutrition Facts panel of the food label. It should include making a specific request to ask people to go to the grocery store and compare the fiber content on the food label of several breakfast cereals they like, and then purchase a cereal that contains 20 percent or more of the Daily Value for fiber per serving. This approach was highly effective in transforming apathy into keen interest in nutrition among working and middle-class women attending focus groups sponsored by the National Cancer Institute (1 4). This segment of women gatekeepers, in particular, may be encouraged to begin taking action as they experience more short-term benefits that are meaningful and motivating. To accomplish this, nutrition education and promotion efforts for Poor Eaters should move them from being interested to being convinced that healthful eating is meaningful and relevant to them. Summary The most effective ways to reach these women gatekeepers by segment is as follows: 1. Better Eaters: Offer new tips that can be added to their current actions for eating healthfully. 2. Fair Eaters: Insert frequent environmental cues to eating healthfully that will appeal to their interest in taste and convenience. 3. Poor Eaters: Establish relevance by identifying ways to appeal immediately to this audience that are consistent with their lifestyle and cultural context. Family Economics and Nutrition Review These fmdings are consistent with those of authors reviewing nutrition education for adults (5). In their review of successful nutrition education interventions for adults, the authors suggested nutrition education communication and strategies in programs that • Are ongoing and multifaceted; • Use mass media to increase awareness and enhance motivation; • Tailor strategies based on formative audience research; • Use motivational messages and educational strategies; and • Employ a behaviorally focused approach that is based on personal factors, behavioral capabilities, and environmental factors. The results of this study suggest that nutrition educators can apply the same segmentation methods used by social marketers and health communicators. It can be expected that doing so would allow them to make the most effective use of resources and to increase program efficiency. We suggest that with a greater understanding of applicable target segments, nutrition educators, policymakers, and other information multipliers will be betterpositioned to improve the diets of Americans. 2001 Vol. 13 No. 1 References I. Ajzen, I. and Madden, T.J. 1986. Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology 22:453-4 72. 2. Andreasen, A.R. 1995. Marketing Social Change. Jossey-Bass Publishers, San Francisco, CA. 3. Bandura, A. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice-Hall, Englewood Cliffs, NJ. 4. Bowman, S., Lino, M., Gerrior, S., and Basiotis, P. 1998. The Healthy Eating Index 1994-96. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion. CNPP-5. 5. Contento, 1., Balch, G.I., Bronner, Y.L., Paige, D.M., Lytle, L.A., Maloney, S.K., Olson, C.M., and Swadener, S.S. 1995. Nutrition education for adults. Journal of Nutrition Education 27(6):312-328. 6. Curry, S.J., Krista!, A.R., and Bowen, D.J. 1992. An application of the stage model of behavior change to dietary fat reduction. Health Education Research 7:97-105. 7. Davis, C. and Saltos, E. 1996. The Dietary Guidelines for Americans- Past, present, future. Family Economics and Nutrition Review 9{2):4-13. 8. Eisner, E., Loughrey, K., and Hairston, B. 1995. Focus Groups with African Americans on Nutrition and Cancer. National Cancer Institute, Bethesda, MD. 9. Federal Program Update. 1996. Nutrition and physical activity communications at the Centers for Disease Control and Prevention. Social Marketing Quarterly 3: 14-15. 10. Frazao, E. 1996. The American diet: A costly health problem. FoodReview 19(1):2-6. II . How Do Americans ' Diets Measure Up To The 1995 Dietary Guidelines For Americans? July 1998 (7). U.S. Department of Agriculture, Food Surveys Research Group. 12. Kover, A. 1995. Copywriters' implicit theories of communication: An exploration. Journal of Consumer Research 21 :596-611. 13. Lefebvre, R.C., Doner, L., Johnston, C., Loughrey, K., Balch, G.I., and Sutton, S.M. 1995. Use of database marketing and consumer-based health communication in message design: An example from the Office of Cancer Communications "5 A Day for Better Health" program. In E. Maibach and R.L. Parrott (Eds.), Designing Health Messages: Approaches from Communication Theory and Public Health Practice (pp. 217-246). Sage, Thousand Oaks, CA. 14. Loughrey, K., Eisner, E., Doner, L., and Weinberg, L. 1994. Exploring the Nutrition Facts Food Label As a Too/for Educating Consumers about NCI's Dietary Guidelines, A Focus Group Study. National Cancer Institute. 15. Loughrey, K.A., Balch, G.I., Lefebvre, R.C., Doner, L., Johnston, C., Eisner, E., and Hadley, L. 1997. Bringing 5 A Day consumers into focus: The qualitative use of consumer research to guide strategic decision making. Journal of Nutrition Education 29(4): 172-177. 13 16. Maibach, E. W., Maxfield, A., Ladin, K., and Slate, M. 1996. Translating health psychology into effective health communication: The American Healthstyles Audience Segmentation Project. Journal of Health Psychology I :261-277. 17. Perry, C.L., Baranowski, T., and Parcel, G.S. 1990. How individuals, environments, and health behavior interact: Social learning theory. ln K. Glanz, F.M. Lewis, F., and B.K. Rimer (Eds.), Health Behavior and Health Education Theory, Research, and Practice (pp. 161-186). Jossey-Bass Publishers, San Francisco, CA. 18. Prochaska, J.O. and DiClemente, C.C. 1982. Transtheoretical therapy: Toward a more integrative model of change. Psychotherapy: Theory, Research and Practice 19:276-288. 19. Prochaska, J.O. and DiClemente, C. C. 1985. Common processes of self-change in smoking, weight control, and psychological distress. ln S. Shiffman and T. Wills (Eds.). Coping and Substance Use (pp. 345-363). Academic Press, Orlando, FL. 20. Research Triangle Institute. 1998 (July 20). Why SUDAAN? [On-line]. Available: http: //www. rti.org/patents/sudaanlwhy sudaan.html. 21 . Slater, M.D. 1995. Choosing audience segmentation strategies and methods for health communication. In E. Maibach and R.L. Parrott (Eds.). Designing Health Messages: Approaches From Communication Theory and Public Health Practice (pp. 186-198). Sage, Thousand Oaks, CA. 22. Sutton, S.M., Layden, W., and Haven, J. 1996. Dietary guidance and nutrition promotion: USDA's renewed vision of nutrition education. Family Economics and Nutrition Review 9(2): 14-21 . 23. U.S. Department of Agriculture and U.S. Department of Health and Human Services. 2000. Nutrition and Your Health: Dietary Guidelines for Americans (5th ed.). U.S. Department of Agriculture. Home and Garden Bulletin No. 232. 24. U.S. Department of Health and Human Services. 2000. Healthy People 2010. Washington, DC. 14 Family Economics and Nutrition Review Anna Maria Siega-Riz, PhD Department of Nutrition School of Public Health University of North Carolina at Chapel Hill, NC, USA Claude Cavadini Nestle Research Center Lausanne, Switzerland Barry M. Popkin, PhD Department of Nutrition School of Public Health University of North Carolina at Chapel Hill, NC, USA 2001 Vol. 13 No. 1 U.S. Teens and the Nutrient Contribution and Differences of Their Selected Meal Patterns We examined the nutrient contribution of foods consumed at breakfast, brunch, lunch, dinner, and snacks, as well as the types of foods consumed on those occasions, by adolescents (n= 1,31 0) participating in the 1989-91 Continuing Survey of Food Intakes by Individuals. Descriptive statistics were generated, using weights and taking into account sample design effects, to examine the consistency of their meal patterns: Consistent, moderately consistent, and inconsistent. Results showed that for individuals with an inconsistent meal pattern, dinner provided half of the day's energy and total snacks provided over one-fifth, an equivalent of one meal for others. Most nutrients studied (fat, protein, calcium, and iron) followed the same pattern as energy. Age differences were noted : 15- to 18-year-olds were more likely to have inconsistent patterns. The types of foods consumed also differed by meal pattern. Both increasing the consistency in the number of meals consumed, as well as improving food-selection behaviors, may serve as possible interventions to improve the diets of adolescents. A dolescence is a period of great transitions. Nutrient require ments are increased from childhood because of physical growth, and behaviors acquired during this period persist into adulthood (1,11, 17,22). While many subsets of adolescents engage in behaviors that have wide public health attention, some adolescents may also follow pathways of poor food choices and reduced physical activity-both of which can also have deleterious effects on health (10,25). Among the health consequences of following these pathways have been rapid increases in obesity and adult-onset diabetes (1 3,23,26). Members of this age group are influenced strongly by their peers, the media, and family situation and less by their knowledge of risky behaviors (6,21,22). Skipping meals is a common practice among adolescents: about 20 percent do not eat breakfast, and about half as many do not eat lunch (3,5,10,18,19). Skipping meals may lead to more snacking; for those who do not view skipping meals as a method of weight loss, snacks often compensate for missed calories and other key nutrients. The literature indicates that, on average, most children and adolescents average four eating occasions a day, with an upper range of 13 occasions among Mexican children who consumed as much as 45 percent of their energy from snacks (4,7,8). Research on the meal patterns of U.S. adolescents showed that most consume at least two meals (plus or minus snacks) on a consistent basis while some follow a highly inconsistent meal pattern: one meal and/or snacks all day (18). 15 Compared with adolescents with inconsistent meal patterns, those with consistent meal patterns consumed a diet that was adequate in calories and more nutrient dense (with respect to calcium, iron, vitamin E, and fiber) (1 1). Our study examines in more detail the types of food consumed by adolescents at each eating occasion and the nutrient contributions provided by each eating occasion to adolescents' total daily intakes. This study is unique: we examine snacking behaviors by using a nationally representative sample, and we determine the nutrient contributions of snacks. Previous studies have examined only the nutrient density of meals versus snacks without considering their contribution to the total diet, or previous studies have used very small samples to examine this research question (2,16). Methods Survey Design Food consumption data were provided by the 1989-91 Continuing Survey of Food Intakes by Individuals (CSFII), a survey conducted by the U.S. Department of Agriculture's Agriculture Research Service. A nationally representative sample was collected by using a multistage, stratified sample design of the 48 coterminous States and Washington, DC. Data were collected in four waves during each year: one in each season, between April 1989 and May 1991. In each wave, a different sample of participants was selected. The total number of participants in all age groups sampled was 15,192. Dietary data were collected for each individual in selected households. Using a 24-hour recall and two 1-day food records, individuals reported 3 consecutive days of intake. The female head of the household reported dietary intake for individuals less than 12 years 16 old. We were interested in the eating patterns of adolescents, thus our analysis was restricted to 11- to 18- year-olds who reported 3 days of dietary intake (n= 1,31 0). The classification of individuals into meal-pattern categories did not differ between 11- and 12-year-olds, and the differences in nutrient composition of reported intakes of 11- and 12-year-olds were similar in magnitude to the differences between 12- and 13-year-olds. Therefore, 11-year-olds were included in the analysis despite differences in methods of data collection for dietary intake. Variables Meal Paffems Survey data include descriptors of eating occasion (breakfast, lunch, dinner, supper, snack, brunch, and extended consumption) as well as the time of day each food was consumed. To identify meal patterns, we first developed clear and invariable terminology for eating occasions: Breakfast, lunch, brunch, dinner, or snack. Respondents provided the name for each meal. When respondents reported consuming either supper or dinner, the eating occasion was designated as dinner; when the respondent reported consuming both supper and dinner, dinner was designated as lunch and supper designated as dinner. This categorization was based on analysis of the data, which indicated that dinner was consumed primarily as an evening meal (85 to 87 percent between 4 and 8 p.m.). When both supper and dinner were consumed, dinner was the midday meal (56 to 69 percent betwee!l 11 a.m. and 3 p.m.) and supper was the evening meal (70 to 81 percent between 4 and 8 p.m.). Eating occasions for 1.3 percent of foods were unknown or identified as extended consumption and therefore not included in our analysis. Three meal-pattern categories were created based on their ability to provide meaningful comparison of eating behaviors: Consistent, moderately consistent, and inconsistent. These categories are mutually exclusive and include all possible combinations of eating occasions. Respondents with a consistent meal pattern (n=538) consumed two or three meals (plus or minus snacks) on all 3 days of reported intake. Those with a moderately consistent meal pattern (n=726) consumed two or three meals (plus or minus snacks) on 2 of the 3 days of reported intake. And respondents with an inconsistent meal pattern (n=46) consumed only one meal (plus or minus snacks) or snacks only on all 3 days of reported intake. Personal, Household, and Demographic Charaderistics Population characteristics available directly from the CSFII were age, gender, race, region of residence, supplement use, school attendance, educational and employment status of the female head of the household, income status, and household size. Our derived variables were consumption of school-based meals and single- versus dual-parent households. Respondents who reported consuming at least one school-based lunch per week were classified as consumers of school lunch. This method was repeated for school breakfast. Classification as a single- or dual-parent household was based on the presence of a male or female head of household or female and male heads of household, respectively. Nutrient and Food-Grouping System The nutrient database was provided by USDA Survey Nutrient Data Base, Release #7 and was developed for the 1991 CSFII. For this analysis, we used nutrient information provided as the total average intake or as the average percentage of the Recommended Dietary Allowances (RDA) for all nutrients (1 2) consumed over 3 days. Family Economics and Nutrition Review The age- and gender-appropriate RDA values were used to calculate the average percentage of the RDA consumed. Grams of food consumed at each eating occasion were calculated by using the University of North Carolina at Chapel Hill food-grouping system. This system disaggregates major USDA food groups into 56 more distinct nutrient-based groups based on the composition of fat and dietary fiber. The University of North Carolina at Chapel Hill 's food-grouping system covered all foods that respondents reported eating (14,1 5). I Statistical Methods We used Student's t test and a chisquare test to compare the sociodemographic characteristics among the groups based on their meal patterns. Statistical testing, however, was not performed on the proportion of the nutrients or the grams of food contributed by each meal. To do so would have required many comparisons, resulting in our having to use a very stringent p value. Hence our analysis is descriptive. The results provide estimates representative of the U.S. population in the coterminous 48 States. We weighted the statistics for nonresponse and corrected the standard errors for the complex multistage design. We used the STATA survey option that allows for the effects of the complex sample design (20). Results Sociodemogra phic Characteristics Forty-one percent of the adolescents had consistent meal patterns; only 4 percent had inconsistent meal patterns (table 1). The 15- to 18-year-olds were more likely to have inconsistent meal 1This information is available upon request. 2001 Vol. 13 No. I Table 1 . Descriptive characteristics of 15- to 18-year-olds and their households, by meal-pattern category, 1989-91 CSFII Meal pattern Sociodemographic Moderately characteristic Consistent consistent Inconsistent Sample 538 726 46 Percent Female 47.0 52.61 60.91,2 Black 14.1 19.01 23.91,2 Attends school 93.1 87.21 75.61 Single-parent household 27.3 31.41 34.81 Female head of household attended college 34.9 31.1 1 27.31,2 Female head of household has < 12 years of education 26.9 31.11 31.81 Region Northeast 18.0 17.2 10.91.2 Midwest 30.9 26.51 28.3 South 31.2 35.41 34.8 West 19.9 20.9 26.1 1,2 Mean (± S.D.) Percentage of poverty 329 (264) 326 (249) 330 (209) Household size 4.7 (1 .9) 4.2 (1.4)3 4.2 (1 .5)3'4 1 Significantly different from the consistent meal pattern, chi -square analysis, p < 0 .05. 2Significontly different from moderately consistent meal pattern, chi-square analysis, p < 0.05. 3Significantly different from consistent meal pattern, weighted t test, p < 0.05. 4Significantly different from moderately consistent meal pattern, weighted ttest, p < 0.05. patterns. The consistent meal-pattern category had a higher percentage of respondents who were male, white, and who attended school. Adolescents with a consistent meal pattern were less likely to be from a single-parent household and more likely to be from a household in which the female head attended college. Neither the mean percentage of poverty nor years of education of the female bead of household differed significantly by meal-pattern category. Nutrient Profiles Based on Adolescents' Meal Pattern Adolescents with a consistent or moderately consistent meal pattern consumed 37 to 38 percent of their total energy from dinner. For adoles-cents in the inconsistent group, 43 percent of their total energy was consumed at dinner. Even more important is the difference in the role of snacks in their diet. Snacks comprised about 23 percent of the total day's energy for those following an inconsistent meal pattern but only 11 to 16 percent for those following the other two meal patterns. In total, the dinner meal and total snacks together provided more than two-thirds of the day's total energy for adolescents with an inconsistent meal pattern. Differences by age group were noted (figs. 1 and 3). For instance, 11- to 14-year-olds in the inconsistent group obtained 24 percent of energy from lunch while older adolescents in the same group obtained 9 percent of 17 Figure 1. Distribution of energy obtained, by meal-pattern groups of 11-to 18-year-olds, 1989-91 CSFII Meal-pattern groups Consistent Cinner Total 38%-snacks 11% Brlllch 1% C3in8%ne r Moderately consistent 25% Total sna~s 16% Mlalng 1% Inconsistent Total lniOkl 1%'11 Figure 2. Distribution of energy obtained, by meal-pattern groups of 11-to 14-year-olds, 1989-91 CSFII Meal-pattern groups Consistent Moderately consistent Total Inconsistent Total .... ks 11% Lunoh 24'4 Figure 3. Distribution of energy obtained, by meal-pattern groups of 15-to 18-year-o/ds, 1989-91 CSFII Meal-pattern groups 18 Consistent Total ~~lllllllllfll~sn~a~s 12'1. 0.01% Breakfast 20'1. Moderately consistent Inconsistent Total Total 25% Family Economics and Nutrition Review energy from lunch. For II- to 14-yearolds with inconsistent meal patterns, breakfast provided 5 percent of energy, compared with almost twice that amount for the older adolescents. Brunch was more common for older adolescents with inconsistent meal patterns. Among older adolescents (15- to 18-year-olds) following an inconsistent meal pattern, dinner and snacks were far more important than they were for younger adolescents (II- to 14-year-olds). Most nutrients follow the same pattern as that for energy (table 2). Breakfast provided the same proportion of nutrients for adolescents with consistent and moderately consistent meal patterns; whereas, lunch appeared to have lower proportions of fat, protein, carbohydrates, calcium, fiber, and sodium for adolescents with moderately consistent meal patterns. The proportions of nutrients from brunch were very low (no more than 3.5 percent) for adolescents with the consistent and moderately consistent meal patterns but closer to I 0 percent for their counterparts with inconsistent meal patterns. Dinner provided similar proportions of nutrients for all three groups. For adolescents with inconsistent meal patterns, the proportion of nutrients provided by total snacks was nearly double the nutrients provided to adolescents with consistent meal patterns. The proportion of nutrients corning from snacks for the moderately consistent group falls between those of the other two groups. Food Consumed Based on Adolescents' Meal Patterns Interesting differences were noted in the types of foods consumed at each eating occasion across meal-pattern groups. At breakfast, adolescents with a consistent meal pattern, compared with adolescents in the other groups, had 200 I Vol. 13 No. I higher per capita consumption of both low- and medium-fat milk, egg items, low-fiber breads, cooked and ready-toeat cereals, high-fat desserts, and juices (fig. 4). In contrast, adolescents with moderate and inconsistent meal patterns consumed more soft drinks. At lunch, adolescents with consistent meal patterns consumed more milk and higher amounts of total poultry, high-fat desserts, vegetables, fruits, and high-fat grain-based mixed dishes (pizza and macaroni and cheese, etc.), compared with other adolescents (fig. 5). There was no difference in beef! pork consumption between adolescents with consistent and moderately consistent meal patterns, which were, however, higher than that for adolescents with inconsistent meal patterns. The inconsistent and consistent groups had the same quantity of high-fat potato consumption. For dinner, teens with an inconsistent meal pattern had higher intakes of poultry, green/orange vegetables, highfat grain-based mixed dishes, high-fat breads, and soft drinks and a lower intake of low- and medium-fat milk, soy and legumes, and fruits, compared with their other adolescent counterparts (fig. 6). In contrast, snacks for the inconsistent group contained more grams per capita of milk items, in particular medium-fat milk items (whole milk and milk shakes) and soft drinks than was the case for the consistent group (fig. 7). All three groups of adolescents had similar intakes of fruits, high-fat desserts, high-fat salty snack items (chips, salty crackers, etc.) and high-fat grain-based mixed dishes. Patterns within each group revealed that the amount of soft drink consumed per capita at lunch, dinner, and total snacks was higher than the amount of milk consumed (figs. 4-7). High-fat, low-fiber bread is more commonly eaten at breakfast, compared with other bread options available. Low-fiber In total, the dinner meal and total snacks together provided more than twothirds of the day's total energy for adolescents with an inconsistent meal pattern. 19 Table 2. Proportion of nutrients provided by each meal among 11- to 18-year-olds, by meal-pattem consumption 1 Consistent meal pattern Moderately consistent meal pattern Inconsistent meal pattern Nutrient Mean Standard deviation Mean Standard deviation Mean Standard deviation Breakfast Energy 19.23 9.69 18.24 10.62 7.94 8.51 Fat 16.07 11.76 15.66 11 .56 6.31 7.94 Saturated fat 17.92 12.87 17.03 12.20 7.67 8.79 Protein 17.27 9.95 16.15 10.60 9.63 11.86 Carbohydrate 21.95 10.28 20.81 11.93 8.70 9.38 Calcium 28.94 16.53 26.31 16.81 15.81 18.04 Cholesterol 23.35 20.16 21 .35 20.82 10.46 17.45 Iron 29.24 17.26 27.47 19.01 19.72 25.64 Folate 39.11 19.16 35.90 22.16 22.66 25.07 Zinc 20.70 13.16 18.63 15.20 10.51 14.33 Fiber 14.68 10.40 16.51 13.05 7.92 12.93 Sodium 16.23 9.54 16.17 10.40 8.86 10.52 Brunch Energy 0.47 3.48 2.11 5.80 8.58 13.69 Fat 0.51 3.92 2.44 7.85 9.82 16.16 Saturated fat 0.58 3.93 2.47 7.88 9.07 15.41 Protein 0.45 3.16 2.05 6.84 7.02 11.39 Carbohydrate 0.46 3.60 2.04 5.39 7.97 12.99 Calcium 0.60 3.76 2.26 7.36 8.05 16.38 Cholesterol 0.39 2.88 3.50 12.84 7.97 12.93 Iron 0.36 2.52 2.12 6.87 8.00 12.84 Folate 0.40 2.69 2.41 7.76 9.30 16.49 Zinc 0.43 3.04 1.87 5.76 8.65 15.52 Fiber 0.57 5.12 2.13 6.84 8.14 15.44 Sodium 0.39 2.88 2.11 6.30 9.89 17.09 Lunch Energy 31.22 11 .64 24 .52 12.58 11.94 18.98 Fat 33.29 14.11 26.10 14.56 12.52 20.91 Saturated fat 32.43 14.53 26.33 15.62 11 .94 19.99 Protein 30.46 12.04 24.69 14.14 10.32 18.14 Carbohydrate 30.03 12.02 23.48 12.39 12.39 19.05 Calcium 30.46 15.74 24 .82 16.11 10.32 19.26 Cholesterol 26.65 14.67 22.49 17.68 10.41 18.49 Iron 26.36 12.03 21 .56 13.00 9.54 16.59 Folate 22.88 12.70 19.90 14.38 8.88 18.33 Zinc 27.74 12.99 23.75 14.88 10.42 18.25 Fiber 33.16 15.87 25.26 14.25 12.65 22.14 Sodium 32.26 12.70 25.06 14.12 11.52 20.20 1 The percentage may not total to 1 00 for each nutrient because of rounding and the sm~ll percentage of foods with a missing eating-occasion classification. (Continued) 20 Family Economics and Nutrition Review Table 2. Proportion of nutrienfs provided by each meal among 11- to 18-year-olds, by meal-pattem consumption 1 (continued) Consistent meal ~attern Moderately consistent meal ~attern Inconsistent meal pattern Nutrient Mean Standard deviation Mean Standard deviation Mean Standard deviation Dinner Energy 37.72 10.53 38.01 14.60 43.21 26.07 Fat 40.17 13.68 39.79 16.77 46.58 26.58 Saturated fat 38.54 14.52 37.87 17.10 44.04 28.06 Protein 45.23 12.33 45.90 16.52 55.07 25.52 Carbohydrate 33.61 10.92 33.92 14.65 37.49 26.89 Calcium 29.86 14.20 30.84 17.28 42.09 29.05 Cholesterol 43.29 18.86 40.97 20.76 49.50 28.39 Iran 36.81 12.84 37.00 17.46 44 .53 30.34 Folate 29.93 13.59 30.53 16.87 38.64 26.11 Zinc 43.80 14.72 43.86 18.95 49.59 28.25 Fiber 41.84 15.89 41 .61 18.22 46.14 27.62 Sodium 44.83 12.15 44.52 16.40 52.51 26.52 Total snacks Energy 10.92 9.69 15.50 13.74 23.16 19.02 Fat 9.65 10.06 14.40 15.01 21.41 20.34 Saturated fat 10.16 10.94 14.78 15.44 23.38 21 .52 Protein 6.38 7.17 9.85 11 .30 15.36 15.93 Carbohydrate 13.35 11 .53 18.06 15.16 26.91 20.41 Calcium 9.71 11 .97 14.24 15.90 19.52 17.71 Cholesterol 6.05 8.59 10.09 13.08 17.83 21.74 Iron 7.06 8.62 10.45 11.87 15.82 17.82 Folate 7.48 9.93 9.79 12.26 18.75 19.36 Zinc 7.09 8.33 10.53 11 .81 17.37 17.92 Fiber 9.39 10.96 13.11 13.41 22.09 19.26 Sodium 6.12 6.92 10.74 13.18 14.26 14.47 1 The percentage may not total to 100 for each nutrient because of rounding and the small percentage of foods with a missing eating-occasion classification. cereal is consumed more than high fiber, and citrus juices are consumed more than noncitrus juices or fruit drinks at breakfast. At lunch, medium fat beef/pork and poultry were eaten more than the low- or high-fat option. The grams per capita for luncheon meats was equally distributed among each fat option. In contrast to breakfast, the low-fat, low-fiber bread option and fruit drinks were eaten in greater amounts at lunch for the consistent and moderately consistent 2001 Vol. 13 No. I groups. And for all three groups, there was a higher per capita consumption of the high-fat versus the low-fat of grain-based mixed dishes. The type of bread consumed at the dinner meal was similar to that seen at breakfast. And the higher fat version of grain-based meals was once again consumed more than the low-fat version at dinner for all three groups. Patterns within groups for the different types of foods consumed as snacks were similar. Discussion Dietary intake patterns of U.S. adolescents are poor. Skipping meals, excessive snacking, and consumption of excessive high-fat, poor nutritionally dense foods are many of the issues raised in the literature. However, few studies have used nationally representative samples to examine the meal and food patterns of U.S. adolescents. This study highlights the large variation in eating patterns among U.S. adolescents. 21 Figure 4. Grams consumed at breakfast by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups1 Figure 5. Grams consumed at lunch by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups1 Breakfast food items (grams) Low-lot milk Medium-lot milk Eggs and egg dishes Low-lot/low-fiber bread Low-lot/high-fiber bread High-fat/low-fiber bread High-fat desserts Posta, rice, and cooked cereals Ready-to-eat cereals Citrus fruits and juices Other fruits and juices Sugars and jellies Coffee and teo Soft drinks Fruit drinks 0 20 40 60 • Consistent 70 68 80 85 8 100 120 Lunch food items (grams) Low-fat beef and pork Medium-fat beef and pork Low-fat poultry Medium-fat poultry Low-/medium-fat luncheon meats/hot dogs High-fat luncheon meats/ hot dogs High-fat desserts High-fat grain-based mixed dishes Citrus fruits and juices Other fruits/juices High-fat potatoes Vegetables Soft drinks 88 4 0 20 40 60 80 100 0 Moderately consistent • Inconsistent • Consistent D Moderately consistent • Inconsistent 1 University of North Carolina at Chapel Hill food-grouping system. 1 University of North Carolina at Chapel Hill food-grouping system. In particular, we show that teens differ markedly by the proportion of food intake from each meal and the types of foods eaten, based on consistent, moderately consistent, or inconsistent meal patterns. Regarding snacks, our results differ from the few published studies on this topic. Our study fmds that snacks contribute much less to the total diet than reported previously. For most adolescents (97 percent), meals contribute, on average, 20 to 40 percent 22 of the total day's energy, compared with 10 to 15 percent contributed by snacks. One study has found that about 25 to 33 percent of the total day's energy comes from snacks (16). Other publi~hed studies have focused more on the frequency of snacking and the snack foods adolescents like to eat (4,5,8). Our study shows that for all the adolescents, a higher proportion of the total day's intake of fat is consumed at dinner. Otherwise, meals and snacks provide similar proportions of the other macronutrients. Our results regarding macronutrients offer a different view of examining macronutrient intake; others who have examined the nutrient density of meals and snacks have found that meals are higher in fat and lower in carbohydrates than are snacks (8). The nutrient contribution of snacks is more significant for those adolescents following an inconsistent meal pattern, compared with adolescents following Family Economics and Nutrition Review Figure 6. Grams consumed at dinner by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups' Figure 7. Grams consumed at snack by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups' Dinner food items (grams) Soy and legumes High-fat/ low-fiber bread Green and orange vegetables Citrus fruits/ juices Other fruits Diet soft drinks 0 20 5 40 60 80 100 120 • Consistent Snack food items {grams) Low-fat milk Medium-lot milk Low-fat desserts High-fat desserts High-fat, salty snacks High-fat groin-based mix Citrus fruits/ juices Other fruits Candy Fruit drinks Diet soft drinks 0 20 40 60 80 100 120 140 160 180 • Consistent 0 Moderately consistent • Inconsistent D Moderately consistent • Inconsistent 1 University of North Carolina at Chapel Hill food-grouping system. 1 University of North Carolina at Chapel Hill food-grouping system . other meal patterns. This occurs simply by how this group was defmed, those consuming one meal plus or minus snacks on 3 days of intake. The nutrient contribution of snacks for this adolescent group is similar to that reported by Ruxton et al. (1 6) in a study of 7- to 8- year-olds (n=136), from five schools in Scotland). In our study, snacks for the inconsistent group provided more of most nutrients than did breakfast or lunch with the exception of iron and folate, which were higher at breakfast. By using the 1977 Nationwide Food Consumption Survey (NFCS), 2001 Vol. 13 No. I researchers found that for most adolescents, snacks compared with meals contributed significantly more magnesium, calcium, vitamin A, and vitamin C to the diet (2). For the only nutrient on which we overlapped, calcium, this was not found in the 1989-91 USDA survey. Because of the frequency of snacking and the significant proportion of energy and other nutrients that snacks provide adolescents with an inconsistent meal pattern, we believe the nutritional quality of snacks has important implications for the health status of these adolescents. Another nutritionally important issue is adolescents' high intake of soft drinks and lower intake of milk [also noted in the 1977 NFCS survey (9)]. These consumption patterns apply to adolescents regardless of meal-pattern group. Even though adolescents with a consistent meal pattern consume the most milk, their calcium intakes are lower than recommended. Also, adolescents appear to be consuming more high-fat and low-fiber foods than the more healthful alternatives. Consuming more high-fat and low-fiber foods may have serious health consequences (i.e., 23 obesity, osteoporosis, and cardiovascular diseases) if they are consumed in high amounts throughout life. The reasons for the high consumption of these types of foods may be directly related to their source (home vs. awayfrom- home food sources) as well as the taste preferences of adolescents. A Minnesota survey of 900 adolescents reported a strong preference for highfat foods that related to taste appeal despite the health consequences associated with consumption of these foods (21). A limitation of this study is the small sample size for the group with an inconsistent meal pattern. However, there were 27 million adolescents in the United States (24) around the time of this survey. Thus 950,000 U.S. adolescents are represented as following an inconsistent meal pattern for 3 days. In general, adolescents are consuming a large quantity of carbonated beverages and few fruits and vegetables. And for adolescents who follow an inconsistent meal pattern, dinner and snacks provide a disproportionate amount of nutrients. Differences are also noted in food selection: adolescents following an inconsistent meal-pattern group consume more types of fast foods . Both meal-pattern and food-selection behaviors should be used to target future public health messages to adolescents. More research is warranted on the determinants of adolescent eating patterns. Information on the determinants could help guide interventions for changing eatingpattern behaviors noted in this study. 24 Acknowledgments We thank the Nestle Research Center for providing financial support for this study. We thank Dr. Henri Dirren for initiating the collaboration between Nestle and the University of North Carolina at Chapel Hill. We also thank Dan Blanchette, Terri Carson, Claire Zizza, Lynn Igoe, and Frances Dancy for their assistance. Family Economics and Nutrition Review References I. Berenson, G.S., McMahan, C.A., Voors, A.W., Webber, L.S., Srinivasan, S.R., Frank, G.C., Foster, T.A., and Blonde, C.V. 1980. Cardiovascular Risk Factors in Children, the Early Natural History of Atherosclerosis and Essential Hypertension. Oxford University Press, New York. 2. Bigler-Doughten, S. and Jenkins, R.M. 1987. Adolescent snacks: Nutrient density and nutritional contribution to total intake. Journal of the American Dietetic Association 87:1678-1679. 3. Cavadini, C. 1996. Dietary habits in adolescence: Contributions of snacking. In A. Ballabriga (Ed.), Feeding from Toddlers to Adolescence. Nestle Nutrition Workshop Series vol. 37. Lippincott-Raven, Philadelphia. Nestle Nutrition Services, Vevey, Switzerland. 4. Cross, A.T., Babicz, D., and Cushman, L.F. 1994. Snacking patterns among 1800 adults and children. Journal of the American Dietetic Association 94: 1398-1403. 5. Devaney, B.L., Gordon, A.R., and Burghardt, J.A. 1995. Dietary intakes of students. American Journal of Clinical Nutrition 6l(supp 1 ):205S-212S. 6. Friedman, H.L. 1989. The health of adolescents: Beliefs and behavior. Social Science and Medicine 29:309-315. 7. Garcia, S.E., Kaiser, L.L., and Dewey, K.G. 1990. The relationship of eating frequency and caloric density to energy intake among rural Mexican preschool children. European Journal of Clinical Nutrition 44:381-387. 8. Gatenby, S.J. 1997. Eating frequency: Methodological and dietary aspects. British Journal of Nutrition 77(supp l):S7-S20. 9. Guenther, P.M. 1986. Beverages in the diets of American teenagers. Journal of the American Dietetic Association 86:493-499. 10. Lund, E.K., Lee-Finglas, W.E., Southon, S., Gee, J.M., Johnson, I.T., Finglas, P.M., and Wright, A.J. 1992. Dietary fat intake and plasma lipid levels in adolescents. European Journal of Clinical Nutrition 46:857-864. II. Miller, E. C. and Maropis, C. G. 1998. Nutrition and diet-related problems. Primary Care 25:193-210. 12. National Research Council, Subcommittee on the Tenth Edition of the RDAs, Food and Nutrition Board. 1989. Recommended Dietary Allowances (I Oth ed.) National Academy Press, Washington, DC. 13. Pinhas-Hamiel, 0. and Zeitler, P. 1996. Insulin resistance, obesity, and related disorders among black adolescents. Journal of Pediatrics 129(3):319-320. 14. Popkin, B.M., Haines, P.S., and Reidy, K.C. 1989. Food consumption trends of U.S. women: Patterns and determinants between 1977 and 1985. American Journal of Clinical Nutrition 49:1307-1319. 15. Popkin, B.M., Siega-Riz, A.M., and Haines, P.S. 1996. A comparison of dietary trends between racial and socioeconomic groups in the United States. New England Journal of Medicine 335:716-720. 2001 Vol. 13 No.1 25 16. Ruxton, C.H.S., Kirk, T.R., and Belton, N.R. 1996. The contribution of specific dietary patterns to energy and nutrient intakes in 7-8 year old Scottish schoolchildren. III. Snacking habits. Journal of Human Nutrition and Dietetics 9:23-31. 17. Serdula, M.K., Ivery, D., Coates, R.J., Freedman, D.S., Williamson, D.F., and Byers, T. 1993. Do obese children become obese adults? A review of the literature. Preventive Medicine 22:167-177. 18. Siega-Riz, A.M., Popkin, B.M., and Carson, T. 1998. Three squares or mostly snacksWhat do teens really eat?: A sociodemographic study of meal patterns. Journal of Adolescent Health 22:29-36. 19. Siega-Riz, A.M., Popkin, B.M., and Carson, T. 1998. Trends in breakfast consumption for children in the U.S. from 1965-1991. American Journal of Clinical Nutrition 67:748s- 756s. 20. STATA Corporation. 1997. Stata statistical software: Release 5.0. STATA Corporation, College Station, TX. 21. Story, M. and Resnick, M.D. 1986. Adolescents' view on food and nutrition. Journal of Nutrition Education 18: 188-192. 22. Sweeting, H., Anderson, A., and West, P. 1994. Socio-demographic correlates of dietary habits in mid to late adolescence. European Journal of Clinical Nutrition 48:736-748. 23. Troiano, R.P., Flegal, K.M., Kuzmarski, R.J., eta!. 1995. Overweight prevalence and trends for children and adolescents. The National Health and Nutrition Examination Surveys. Archives of Pediatrics and Adolescent Medicine 149: I 085-1091. 24. U.S. Department of Commerce, Economics and Statistics Administration, Bureau of the Census. 1992. 1990 Census of the Population: General Population Characteristics. 25. U.S. Department of Health and Human Services, Public Health Service, Office of Disease Prevention and Health Promotion, Centers for Disease Control, and National Institute on Drug Abuse. 1989. The National Adolescent Student Health Survey: A Report on the Health of Americas Youth. Third Party Publishing Co., Oakland, CA. 26. U.S. Department of Health and Human Services. 1996. Update: Prevalence of overweight among children, adolescents, and adults-United States 1988-94. Morbidity and Mortality Weekly Report 46:199-202. 26 Family Economics and Nutrition Review Rachel K. Johnson, PhD, MPH, RD The University of Vermont Celeste V. Panely, MS, RD The University of Vermont Min Qi Wang, PhD The University of Maryland 2001 Vol. 13 No. 1 Associations Between the Milk Mothers Drink and the Milk Consumed by Their School-Aged Children The declining milk intakes of U.S. children are of concern because milk is the primary source of calcium in children's diets. The aim of th is study was to determine the predictors of milk consumption in U.S. schoolaged children (ages 5- 17) by using dietary intake data from the USDA 1994-95 Continuing Survey of Food Intakes by Individuals (CSFII). Sociodemographic variables, type of milk consumed (skim, 1 %, 2%, whole, or none), and mothers' milk intake (type and amount) were examined as possible predictors. The sample consisted of 1,303 CSFII participants. Sample weights were applied to allow for generalizations to the entire U.S. school-aged population. Children's average milk intake was 300.4 grams per day. For every gram of milk a mother consumed, her child's intake increased by 0 .64 grams. Two percent milk was the most commonly consumed milk among the children. For each type of milk consumed by mothers, children were at least 30 times more likely to drink that same type. The strong association between the milk consumed by mothers and the amount and type of milk consumed by U.S. school-aged children should be considered when designing intervention programs aimed at increasing children 's milk intake. E vidence suggests that attainment of peak bone mass by early adulthood may be the most effective protection against osteoporotic fractures later in life (23). Throughout the developmental years, adequate calcium intake is essential to support bone growth (16). Substantial evidence exists linking higher calcium intakes with improved skeletal health in children (2,3,16,21,23,30). Data from the U.S. Department of Agriculture's (USDA) nationwide food consumption surveys reveal that most U.S. schoolaged children have calcium intakes that are below recommended levels (4). Calcium intake is especially problematic for girls, with 59 percent ages 6-11 and 86 percent ages 12-18 not meeting recommendations (4) . Milk and dairy products are the primary source of calcium in children 's diets (8). Johnson and colleagues found that in a large sample of school-aged children, on average, only those children who consumed milk at the noon meal met their daily requirement for calcium (1 5). Rising consumption of soft drinks has been shown to have a negative effect on calcium intake among children and adolescents by 27 competing with milk as a preferred beverage (9). On the other hand, whole and 2% milk are leading sources of fat and saturated fat in the diets of U.S. children (33). USDA food consumption survey data indicate that for children in all age groups, mean total and saturated fat intakes exceed the recommended levels (4). Because milk is an important contributor of both calcium and fat in the diets of children, it is important to identify the predictors of children's milk intake (both type and amount). The aim of this study was to identify predictors of U.S. school-aged children's milk intake. Familial aggregation studies show similarities in nutrient intake between parents (especially mothers) and their children (26). Hence, milk consumption patterns of mothers were included, along with sociodemographic variables, in the research model as possible predictors of children's milk intake. Findings from this study will assist nutrition policymakers, school nutrition personnel, school administrators, nutrition educators, and parents in developing appropriate intervention strategies to address the problem of children's declining milk consumption. Methods Sample The research sample was obtained from the 1994-95 USDA Continuing Survey of Food Intakes by Individuals (CSFII). The CSFII is a continuing component of the USDA Nationwide Food Consumption Survey. The surveys provide data on demographics as well as dietary intake for a nationally representative sample of noninstitutionalized persons residing in the United States. The 1994-95 survey included data on the food and nutrient intakes of 5,598 individuals. The response rate of the 28 survey was 80 percent for Day I dietary intake data and 76 percent for Day 2 (4). These response rates are acceptable by research standards (7) . Trained interviewers used the multiplepass 24-hour recall method to collect 2 days of dietary intake data from each respondent. The multiple-pass 24-hour recall method has been validated as an accurate measure of children's dietary intake (11). All children ages 5 to 17 years with 2 complete days of dietary intake data (N= 1 ,303) and their mothers were included in this study. Study Variables The study investigated predictors of both the amount and type (skim, 1%, 2%, whole, or none) of milk consumed by U.S. school-aged children. The following sociodemographic variables were assessed as possible predictors: Child gender, age, and race; household income; geographic region; urbanization; and mother's age, education, and occupation. Participation in the USDA Food Stamp Program and participation in the USDA national school lunch and school breakfast programs were also included as possible predictors of a child's consumption of milk. Milk is required to be served in the national school lunch and school breakfast programs (5) . Mothers' milk consumption patterns (both type and amount) were included as potential predictors. A mother's nutrient intake has been shown to influence her child's nutrient intake (26) . In addition, studies by Pelletier and colleagues indicated that.among adult milk drinkers, consumption of lower fat versions of milk (I% and skim) was associated with increased average daily milk consumption (2 7). If the same is true for children, promotion of I% and skim milk in this population could have a positive influence on calcium intake. The dependent variables in the analysis were "Child Milk Amount" and "Child Milk Type." Child Milk Amount was defined as the 2-day mean intake in grams of fluid milk consumed by the sample child. The 7,250 food codes in the CSFII database were searched, and all codes whose primary ingredient was fluid cows' milk were included. Items such as flavored milk, evaporated milk, dry reconstituted milk, eggnog, and milk shakes were included. However, items such as flavored drinks (e.g., Yoo-hoo®l, canned meal replacements (e.g., Instant Breakfast®l, and infant formulas were excluded. Child Milk Type was defined as the type of milk (skim, 1%, 2%, whole, or none) most often consumed by the sample child. The CSFII food codes were searched and all fluid milks were grouped into one of the four categories: Skim, I%, 2%, or whole. For example: "milk, chocolate, skim milk based" was categorized as skim; "milk, dry, reconstituted, whole" was categorized as whole. The category consumed in the greatest quantity in grams over 2 days by each sample child was considered the Child Milk Type. Statistical Analysis The Statistical Export and Tabulation System (SETS) software and the Statistical Analysis System (SAS) were used to format and recode the data for statistical analysis. Statistical significance was set at p S 0.05 for all analyses. To compensate for variable probabilities of selection, differential nonresponse rates, and sampling frame considerations, we applied sample weights in both the descriptive and comparative analyses. The Survey Data Analysis System (SUDAAN) was used to weight the sample, compute variances, and run the statistical procedures. Applying sample weights aiiows the findings to be generalized to the entire U.S. population of school-aged children. Analysis of variance and analysis of Family Economics and Nutrition Review covariance were used to determine both the bivariate and multivariate effect of each independent variable on the dependent variable, Child Milk Amount. Only those independent variables that were significant at the bivariate level were included in the final multivariate model. Chi-square statistics were used to identify independent variables associated bivariately with Child Milk Type. The Multinominal Logistic Model was used for the multivariate analysis of Child Milk Type. As with the Child Milk Amount model, only those independent variables that were significant at the bivariate level were included in the multivariate model. 1 The results of the multinomial model were presented as odds ratios, which describe the change in likelihood of one outcome (e.g., drinking whole milk) versus another outcome (e.g., drinking 2% milk) given a particular characteristic or level of predictor (e.g., being a male compared with being a female) (31). In multinominallogistic models, each outcome (skim, 1%, whole, none) is compared with a reference category, which we determined to be 2% milk-the most common type of milk consumed. Odds ratios greater than 1.0 indicate an increased likelihood of consumption of that type of milk (compared with 2%) for children with that characteristic; whereas an odds ratio of less than 1.0 indicates a lower likelihood of consuming that type of milk (compared with 2%) for children with that characteristic. Both unadjusted and adjusted odds ratios were calculated. 1The Multinominal Logistic Model is an extension of the logistic regression model. While logistic models can only process dichotomous outcome variables, the multinominal model can include outcomes with two or more categories (25). 2001 Vol. 13 No. 1 Table 1. Amount and type of milk consumed1 by children ages 5-17 who provided 2 days of dietary intake data, 1994-95 CSFII Type of milk consumed Percent Mean amount (grams) Skim 11.4 376.62 1% 9.6 407.9 2% 32.0 385.4 Whole 28.4 347.8 None 18.6 0.0 1Two-day mean intake of milk (groms/ day) = 300.4+ 11 .9 . 2There was no association between type (skim, 1%, 2%, whole) and amount of milk consumed. N= 1,303. This allows for the examination of the influence the independent variables have on the dependent variable (Child Milk Type) both before and after the model is adjusted for all the covariates. Any odds ratio with 95 percent confidence intervals that included 1.0 was not considered statistically significant. Results Demographics The unweighted sample of CSFII respondents consisted of 1,303 participants. The children's average age was 11.6 years; the mothers', 39 years. Most of the sample was white, and was divided relatively equally between boys and girls. The sample was geographically diverse and representative of the U.S. population. Most participants resided in suburban areas, and the average yearly household income was about $44,000. The mothers' most common classes of occupation included professionaVtechnical and clericaV sales. Twenty-four percent of the children were eligible to receive free or reduced-price lunches, and 14 percent were eligible to receive free or reduced-price breakfasts. Milk Consumption The 2-day mean milk intake for children was 300.4 grams per day (table 1). Mothers' mean intake was 109.0 grams per day. Of the types of milk consumed by children (skim, l %, 2%, whole, and none), 2% milk was most commonly consumed, followed by whole milk. Two percent milk was also the most commonly consumed type by mothers, followed closely by whole milk. No significant associations were found between the type (skim, I%, 2%, or whole) and amount of milk consumed by children. Predictors of the Amount of Milk Consumed by Children Based on the bivariate analysis, the type and amount of milk consumed by mothers, geographic region, and the child's gender were associated with Child Milk Amount. Hence, these variables were entered into the multivariate model. In this model, the type of milk mothers consumed was not significant; however, geographic region, the child's gender, and the amount of milk mothers consumed each had a significant effect on the amount of milk consumed by children. In the multivariate analysis, children from the Midwest had significantly higher milk intakes than children from 29 For every 1 gram of milk a mother consumed, her child's intake increased by 0.64 grar:ns. 30 Table 2. Amount of milk consumed by children ages 15-17: Analysis of covariance (ANCOVA)1 of significant relationships, 1994-95 CSFII Variable Mothers' milk intake (milk type) Skim 1% 2% Whole None Mothers' milk intake (milk amount, grams) Region Northeast Midwest West South Child's gender Mole Female Beta coefficient (±SE Beta) 1.94±35.1 28.3 ± 28.6 4.3 ± 32.0 29.0 ± 31.7 0.00 ± 0.00 0.64 ± 0.1 11.6 ± 26.9 71.8 ± 35.6 49.3 ± 29.7 0.0 ± 0.0 120.0 ± 16.9 0.0 ± 0.0 P-volue 0.96 0.33 0.89 0.37 <0.001 0.66 0.05 0.10 <0.001 1 F value for overall model = 137.07; P-value for model < .001; Intercept = 145.09. - = No reference category. N=1 ,303. the South (table 2). Boys in the sample consumed 120 grams more milk per day than girls consumed. Maternal milk intake was significantly and positively associated with the amount of milk children consumed. For every 1 gram of milk a mother consumed, her child's intake increased by 0.64 grams. Predictors of the Type of Milk Consumed by Children Of the 12 independent variables, the children's age, gender, and race; geographic region; eligibility for free and reduced-price school lunch and breakfast; mothers' age and level of education; and the amount and type of milk consumed by mothers had a significant bivariate effect on Child Milk Type. Urbanization and participation in the Food Stamp Program were not significant predictors, and were therefore dropped from the multivariate model. In the multivariate model, older children were more likely to drink skim milk or no milk than were younger children (table 3). Children who paid full price for lunch were more likely to drink skim milk, compared with children who were eligible to receive (and presumably received) free school lunch. Children from the Northeast were more likely than children from the South to drink 1% milk; whereas, children from the South were more likely than children from the Midwest to drink whole milk or no milk. Black children were more likely to drink whole milk or no milk than were White children. Girls were twice as likely to drink no milk, compared with boys. The type of milk mothers drank was a very strong predictor of the type of milk children drank. Two percent milk was used as the reference category for Child Milk Type, because this was the type most commonly consumed by the Family Economics and Nutrition Review Table 3. Milk consumed by children ages 5-17: Results of unadjusted and adjusted odds ratios/ 1994-95 CSFII CHILD Age (years) 13-17 9-12 5-8 Race Black White Other Gender Female Mole School lunch None Free Reduced Full School breakfast None Free Reduced Full MOTHER Age (years) 40-60 30-39 20-29 Education College graduate Some college High school Type of milk consumed Skim 1% Whole None 2% Amount of milk consumed (grams) > 360 241-360 121-240 0-120 REGION t. Northeast Midwest West South Unadj OR2 1.8 1.3 1.0 0.1 1.0 0.5 1.4 1.0 1.3 0.1 0.4 1.0 2.7 0.8 0.0 1.0 2.4 1.4 1.0 2.1 2.0 1.0 37.7 4.3 4.6 7.2 1.0 1.0 1.9 0.6 1.0 1.3 0.6 0.7 1.0 Skim Adj OR 2.2 1.6 1.0 0.4 l.O l.O 1.5 1.0 1.0 0.2 1.2 1.0 2.8 8.6 0.4 1.0 0.7 0.8 1.0 2.0 1.5 1.0 30.0 4.7 5.9 7.2 1.0 0.9 1.4 0.6 1.0 1.5 0.8 1.0 1.0 1.2, 3.9 0.9, 2.9 0.1' 1.2 0.3, 3.8 0.9, 2.7 0.4, 2.8 0.1 , 0.4 0.2, 7.4 0.5, 15.5 1.1' 69.8 0.1, 3.4 0. ' 3.9 0.1 , 4.5 0.5, 8.0 0.7, 3.1 9.4, 95.8 0.8, 28 .3 1.5, 23 .4 2.0, 26.1 0.2, 3.2 0.5, 4.3 0.1, 2.4 0 0.7, 3.2 0.3, 2.0 0.4, 2.2 Unadj OR 0.7 0.7 1.0 0.1 l.O 0.7 0.9 1.0 3.6 1.1 0.7 l.O 1.4 1.0 1.3 1.0 2.1 2.3 1.0 3.6 2.2 1.0 3.2 67.2 2.8 3.1 l.O 1.6 1.8 0.9 l.O 4.5 0.8 2.0 1.0 1% Adj OR 1.2 0.9 1.0 0.5 1.0 0.4 1.2 1.0 2.3 1.3 0.9 1.0 0.7 1.9 2.6 1.0 1.7 2.8 1.0 2.5 1.6 1.0 3.7 114 2.2 4.2 1.0 95% Cl 0.6, 2.6 0.4, 1.7 0.2, 1.6 0.1' 1.5 0.7, 2.2 1.2, 4.4 0.3, 4.9 0.2, 4.6 0.1 , 4.2 0.2, 21 0.3, 27 0.4, 7.1 0.7, 11 0.9, 6.8 0.8, 3.3 1.1' 12 31, 416 0.5, 8.5 1.2, 15 1.3 0.3, 6.0 2.8 0.6, 12 1.0 0.3, 2.8 1.0 5.5 1.3, 23 0.9 0.2, 3.3 4.0 1.0, 16 1.0- Unadj OR 0.8 0.8 1.0 6.4 1.0 2.7 0.9 1.0 1.2 2.6 2.0 1.0 3.0 7.3 7.9 1.0 0.4 0.7 1.0 0.1 0.4 1.0 3.7 5.2 50.1 10.8 1.0 1.0 1.3 0.5 1.0 l.l 0.3 0.7 1.0 Whole Adj OR 1.2 1.1 1.0 3.3 1.0 l.7 0.7 1.0 0.9 0.8 1.4 1.0 3.9 3.6 3.1 1.0 0.9 1.5 1.0 0 .3 0 .5 1.0 4 .1 8.2 45.8 13.0 1.0 2.0 1.4 0.9 1.0 1.1 0 .4 0.5 1.0 95% Cl 0.7, 2.1 0.7, 1.7 1.7, 6.4 0.6, 4.6 0.4, 1.1 0.5, 1.7 0.3, 2.6 0.4, 4.2 0.8, 18.5 0.6, 23 .9 0.4, 22.9 0.4, 2.2 0.6, 3.9 0.1' 0.7 0.3, 0.8 1.2, 14.6 2.6, 25.8 17.1 , 122.8 6.0, 28 .1 0.8, 5.1 0.4, 4.3 0.4, 2.3 0.4, 2.8 0.2, 0.7 0.2, 1.3 Unodj OR 3.8 1.0 1.0 3.2 1.0 0.8 2.1 1.0 1.4 0.6 0.8 1.0 1.9 1.8 2.7 1.0 2.7 1.4 1.0 1.2 0.8 1.0 4.5 8.1 7.4 6.9 1.0 0.4 0.2 0.3 1.0 0.8 0.3 0.5 1.0 None Adj OR 3.9 l.O l.O 3.0 1.0 1.3 2.1 1.0 1.0 0.3 0.8 1.0 1.8 3.4 3.0 1.0 1.2 1.0 1.0 1.3 0.8 1.0 4.3 8.6 5.9 3.8 1.0 0.7 0.3 0.5 1.0 0.9 0.4 0.7 1.0 95% Cl 2.2, 7.1 0.5, 1.9 1.2, 7.6 0.4, 4.1 1.1' 4.0 0.6, 1.9 0.1' 1.1 0.3, 2.4 0.4, 8.6 0.4, 26.7 0.4, 21 0.4, 3.8 0.3, 3.0 0.4, 3.8 0.4, 1.4 1.4, 13 3.9, 19 1.8, 19 1.6, 9.2 0.3, 2.0 0.1' 1.1 0.2, 1.3 0.4, 2.1 0.2, 0.8 0.3, 1.7 1Consumption of skim, 1%, whole, or no milk is compared to 2% milk, as 2% is the most commonly consumed milk type among both sample children and mothers. Odds ratios whose confidence limits do not include 1.0 are bolded. 20dds ratios. 3Confidence intervals. -=No reference category. N=l ,303. 2001 Vol. 13 No. 1 31 For each type of milk ... consumed by mothers, children were at least 30 times more likely to drink the some milk type as their mothers. 32 sample. For each type of milk (skim, I%, whole, or none) consumed by mothers, children were at least 30 times more likely to drink the same milk type as their mothers. In addition, the more educated a mother was, the less likely her child was to drink whole milk. The odds ratios for school breakfast, mother milk amount, and mothers' age were not significant; the 95 percent confidence intervals for these variables included or were very close to 1.0. Discussion The findings of this study demonstrated that the amount and type of milk consumed by mothers strongly predicted the amount and type of milk consumed by their school-aged children. This study also demonstrated that differences in children's milk consumption patterns were associated with a number of demographic variables. These included regional differences; differences associated with mothers' level of education; and children's age, gender, and race. Limitations The problem of underreporting of food intake is a concern when interpreting dietary intake data (24) . When food consumption surveys are used to obtain dietary intake data, both adults and adolescents tend to underreport their food intake (22). However, there is agreement that individuals of all ages are prone to exaggerate those foods they perceive to be healthful and to underreport foods that are commonly considered "sin" foods (i.e., foods high in sugar and fat) (22). Milk is generally perceived as a healthful food and was not among those foods most likely to be underreported in the CSFII (18). Hence underreporting was not likely to be a significant problem in this study. Factors Influencing the Amount of Milk Consumed by U.S. School-Aged Children The amount of milk consumed by mothers was associated strongly and positively with the amount of milk consumed by their children. Parents guide and direct children's food choices (17). Wardle and colleagues studied parental influences on children's consumption patterns and found significant mother-child correlations for consumption of dietary fat as well as fruit and vegetable consumption (34). Harper and Sanders observed that children sample unfamiliar food consistently more often when they view their parents partaking of the food (10). Children whose mothers do not drink milk may be less likely to sample milk, perceiving milk as an unfamiliar food. Parental monitoring may also influence children's milk consumption. Research has shown that parental monitoring can have a marked effect on children's food selection (I 7). Researchers interviewed over 50 focus groups with children nationwide regarding the factors that influence their consumption of calcium-rich foods. They discovered that a large percentage of children were neither encouraged nor required by their parents to drink milk at home (36). In our study, other predictors of the amount of milk consumed by U.S. school-aged children included children's gender and region. Compared with the girls, the boys consumed an average of I 28 grams per day more milk. This is an important finding, because girls' calcium intakes are also lower than boys' (4) . Girls' energy needs are typically lower than boys'. These lower energy needs may be reflected in lower intakes of all foods and beverages, including milk. On the Family Economics and Nutrition Review other hand, it is possible that some girls may be restricting their food intake . and eliminating or reducing thetr milk intake to cut calories and fat. Girls may initiate dieting behaviors as early as age 6. In one Ohio study of schoolchildren Grades 1 through 5, close to twice as many girls as boys reported restricting or altering their food intake (1). Adequate calcium int~ke is . especially important for grrls-bemg female is an independent risk factor for developing osteoporosis (8). In our study, differences found in children's milk intake by region of residence are also important. Southern girls have the lowest calcium int~kes, compared with girls in other regions (1 2). This study determined that children in the South also have the lowest milk intakes. In addition, they were more likely than children from other regions to drink no milk at all. Increased milk consumption among children in the South could be influential in improving their calcium intakes. Race did not predict the amount of milk consumed. Lactose maldigestion appears to vary widely among different ethnic and racial groups and in the United States is estimated to be about 15 percent in Whites, 80 percent in African Americans, and 90 percent in Asian Americans (19). However, a dairy-rich diet was found to be well tolerated when fed to AfricanAmerican adolescent girls for 21 days (29) . In this study race did not infl~ence total milk intake. This is consistent with findings that most people with lactose maldigestion are able to tolerate a glass of milk at a meal without developing any significant symptoms (32). 2001 Vol. 13 No. 1 Factors Influencing the Type of Milk Consumed by U.S. School-Aged Children The results of this study also demonstrated that a variety of factors influenced the type (skim, 2%, 1%, whole, none) of milk consumed by U.S. . school-aged children. The type of milk consumed by the mothers was associated strongly with the type of milk consumed by their children. This finding was consistent with results of studies conducted by Fischer and Birch, demonstrating that exposure to a food over time will result in the development of a preference for the food among children (6) . Children who have continued exposure to 1% and skim milk in the home and who observe their mothers consuming these types of milk are likely also to dri~ these types of milk. The type of milk consumed by children can have an effect on total diet quality. Children who drink skim milk come closer to meeting dietary recommendations f~r fat and saturated fat in their total daily diet (1 5,28) . Individuals in all age groups who consume I% and skim milk also consume more fruits and vegetables and less red meat (20). On the other hand, the cross-sectional nature of these data make it difficult to sort out the directionality of the association between the type of milk consumed by mothers and the type consumed by their children. Thus, it is possible that mothers m~y simp_ly drink the type of milk their children hke, and if the children do not like milk, mothers may not buy it just for themselves. Prior studies have shown that mothers' education level is correlated with their children's nutrient intake (13). In our study, mothers with the fewest years of education were more likely to have children who drank whole milk or no milk at all, compared with mothers who were more highly educated. Nutrition information may not be reaching less educated mothers. It is also possible that 1% and skim milk are not as accessible to them. Whole milk is sometimes the only choice available in lower income communities (35). Children from the South (compared with those in other regions) as well as black children (compared with white children) were more likely to drink whole milk or no milk at all. It may be necessary to target the Southern United States for outreach, because children in the South have the highest fat and saturated fat intakes and the lowest calcium intakes of children in all regions in the United States (14). Several other variables were associated with the type of milk consumed by school-aged children. Older children and girls were more likely to dri~ skim milk than were younger children and boys, respectively. Findings als~ showed that children eligible to receive free school lunch were less likely to drink skim milk than were children who paid full price. Beginning in the fall of 1996, schools participating in USDA school nutrition programs were required by law to serve meals that on average meet the dietary guidelines for fat, saturated fat, cholesterol, and sodium (5) . Because it is difficult to meet the dietary guidelines when a meal includes whole milk (1 5), participating schools may now be_ serving and marketing I% and skun milk more vigorously. Further research using future USDA surveys is needed to conftrm this possibility. 33 Implications The findings of our study demonstrate that mothers' milk consumption patterns are potentially strongly associated with the type and amount of milk consumed by U.S. school-aged children. Interventions aimed at increasing children's milk consumption should consider the strong influence of maternal modeling on children's milk intake. Mothers should be encouraged to serve as positive role models for their children by drinking skim or 1% milk regularly. In addition, it becomes apparent that milk promotion campaigns targeting women for prevention of osteoporosis may have a spillover effect of increasing children's milk consumption. Acknowledgments This project was funded by the Vermont Agricultural Experiment Station Hatch Project #VT-NS-00577. 34 References I. Berg, F.M. 1997. Afraid to Eat; Children and Teens in Weight Crisis (2nd ed.). Healthy Weight Publishing Network, Hettinger, North Dakota. 2. Cadogen, J., Eastell, R., and Jones, N. 1997. Milk intake and bone mineral acquisition in adolescent girls: Randomized, controlled intervention trial. BMJ 15: 1255-1260. 3. Chan, G.M., Hoffman, K., and McMurry, M. 1995. Effects of dairy products on bone and body composition in pubertal girls. Journal of Pediatrics 126:551-556. 4. Continuing Survey of Food lntakes by Individuals (CSFII), 1994-95 data set descriptions and data tables: Combined results of the USDA's 1994-95 CSFII. ARS Beltsville Human Nutrition Research Center, Food Surveys Research Group, l996. 5. Eadie, R.E. 1995. Child nutrition programs: School meal initiatives for healthy children; final rule. Federal Register 60(113):31188-31222. 6. Fischer, J.O. and Birch, L.L. 1995. Fat preferences and fat consumption of 3-5 year-old children are related to parental adiposity. Journal of the American Dietetic Association 95:759-764. 7. Fowler, F.J. 1990. Survey Research Methods. Applied Social Research Methods Series, Volume I. Sage Publications, Newbury Park, CA. 8. Guthrie, H.A. and Picciano, M.F. 1995. Human Nutrition. Mosby, Boston. 9. Harnack, L., Stang, J., and Story, M. 1999. Soft drink consumption among U.S. children and adolescents: Nutritional consequences. Journal of the American Dietetic Association 99:436-441 . 10. Harper, L.V. and Sanders, K.M. 1975. The effects of adults on young children's acceptance of unfamiliar foods. Journal of Experimental Child Psychology 20:206-214. II. Johnson, R.K., Driscoll, P., and Goran, M.I. 1996. Comparison of multiple-pass 24- hour recall estimates of energy intake with total energy expenditure determined by the doubly labeled water method in young children. Journal of the American Dietetic Association 96: 1140-1144. 12. Johnson, R.K., Guthrie, H., and Smiciklas-Wright, H. 1994. Characterizing nutrient intakes of children by sociodemographic factors. Public Health Reports I 09(3):414-420. 13. Johnson, R.K., Johnson, D., Harvey, J., and Wang, M. 1994. Dietary quality of the noon-time meal among a large sample of U.S. adolescents. School Food Service Research Review 18(1):2-7. 14. Johnson, R.K., Johnson, D., Wang, M., Smiciklas-Wright, H., and Guthrie, H. 1994. Characterizing nutrient intakes of adolescents by sociodemographic variables. Journal of Adolescent Health 15:149-154. 15. Johnson, R.K., Panely, C.V., and Wang, M.Q. 1998. The association between noontime beverage consumption and the diet quality of school-aged children. Journal of Child Nutrition and Management 22(2):95-1 00. 16. Johnston, C.C., Miller, J.Z., Slemenda, D.W., Reister, T.K., Hui, S., Christian, J.C., and Peacock, M. 1992. Calcium supplementation and increases in bone mineral density in children. The New England Journal of Medicine 327:82-87. Family Economics and Nutrition Review 17. Klesges, R.C., Stein, R.J., and Eck, L.H. 1991. Parental influence on food selection in young children and its relationship to childhood obesity. The American Journal of Clinical Nutrition 53:859-864. 18. Krebs-Smith, S.M., Graubard, B., Cleveland, L., Subar, A., Ballard-Barbash, R., and Kahle, L. 2000. Low energy reporters vs others: A comparison of reported food intakes. European Journal of Clinical Nutrition 54:281 -287. 19. Lactose Intolerance. 1994. National Digestive Diseases Information Clearinghouse, Washington, DC. National Institutes of Health publication 94-2751. 20. Lee, H.C., Gerrior, S.A., and Smith, J.A. 1998. Energy, macronutrient and food intakes in relation to energy compensation in consumers who drink different types of milk. The American Journal of Clinical Nutrition 67:616-623. 21. Lee, W., Leung, S., and Wang, S.H. 1994. Double-blind, controlled calcium supplementation and bone mineral accretion in children accustomed to a low calcium diet. The American Journal of Clinical Nutrition 60:744-750. 22. Livingstone, M.B., Prentice, A.M., and Coward, W.A. 1992. Validation of estimates of energy intake by weighted dietary record and diet history in children and adolescents. The American Journal of Clinical Nutrition 56:29-35. 23. Lloyd, T., Andon, M.B., and Rollings, N. 1993. Calcium supplementation and bone mineral density in adolescent girls. lAMA 270(7):841-844. 24. Mertz, W., Tsui, J.C., and Judd, J.T. 1991 . What are people really eating? The relation between energy intake derived from estimated diet records and intake determined to maintain body weight. The American Journal of Clinical Nutrition 54:28-35. 25. Morel, J.G. 1989. Logistic regression under complex survey designs. Survey Methodology 15:203-223. 26. Oliveria, S.A., Ellison, R.C., and Moore, L.L. 1992. Parent-child relationships in nutrient intake: The Framingham Children 's Study. The American Journal of Clinical Nutrition 56:593-598. 27. Pelletier, D.L., Kendall, A., and Mathios, A. 1996. Lowfat milk promotion: Opportunities created by a new policy environment. Unpublished work, pp. 1-25. Cornell University, Ithaca, NY. 28. Peterson, S. and Sigman-Grant, M. 1997. Impact of adopting lower-fat food choices on nutrient intake of American children. Pediatrics I 00(3):A380. 29. Pribila, B., Hertzler, S.R., Martin, B.R., Weaver, C.M., and Savaiano, D.A. 2000. Improved lactose digestion and intolerance among African-American adolescent girls fed a dairy-rich diet. Journal of the American Dietetic Association I 00:524-528. 30. Sentipal, J.M., Wardlaw, G.M., and Mahan, J. 1991. Influence of calcium intake and growth indexes on vertebral bone mineral density in young females. The American Journal of Clinical Nutrition 54:425-428. 31 . Shah, B.V., Barnwell, B.G., and Bieler, G.S. 1997. SUDAAN Software for the Statistical Analysis of Correlated Data. User's Manual. Research Triangle Institute, NC. 32. Suarez, F.L., Savaiano, D.A., and Levitt, M.D. 1995. A comparison of symptoms after the consumption of milk or lactose-hydrolyzed milk by people with self-reported severe lactose intolerance. The New England Journal of Medicine 333:1-4. 2001 Vol. 13 No. I 35 33. Thompson, F.E. and Dennison, B.A. 1994. Dietary sources of fats and cholesterol in U.S. children aged 2 through 5 years. American Journal of Public Health 84:779-806. 34. Wardle, J., Gibson, L., and Watts, C. 1995. Parental influences on children's nutritional knowledge, diet and risk factor status. Annals of Behavioral Medicine PA16A, S079. 35. Wechsler, H., Basch, C.E., Zybert, P., Lantigua, R., and Shea, S. 1995. The availability of low-fat milk in an inner-city Latino community: Implications for nutrition education. American Journal of Public Health 85(12) : 1690-1692. 36. Wolfe, F.H. Factors influencing the intake of calcium rich foods among adolescents. [On-line). Available: http://cristel.nal.usda.regional.linkpage. Accessed December 13, 1999. 36 Family Economics and Nutrition Review Carol Byrd-Bredbenner, PhD, RD Rutgers, The State University of New Jersey Darlene Grasso, MA, RD Montclair State University 2001 Vol. 13 No. l The Effects of Food Advertising Policy on Televised Nutrient Content Claims and Health Claims This study examined changes in nutrient content and health claims made in televised food advertisements before and after the Federal Trade Commission's 1994 food advertising pol icy, which is predicated on the Nutrition Labeling and Education Act (NLEA). Our sample included 1 05 and 1 08 advertisements broadcast during prime-time in 1992 and 1998, respectively. The rate that nutrient content and health claims were used was low in both years. And none of the advertisements contained diet-disease health claims authorized by the Food and Drug Administration. Although current food advertising policy virtually eliminates deceptive advertisements, it may also limit diet-disease health claims in broadcast media . More flexibility in presenting diet-disease health cla ims in broadcast media advertising could increase the use of such claims and contribute to the goal of NLEA to educate consumers. T he decision to purchase a food is influenced by many factors, one of which is advertising (7,8,36,43). Advertisements traditionally promoted foods and beverages by featuring mainly sensory qualities, convenience, and economic factors (10,44). In recent years some of these advertisements have tried to influence consumer-purchasing decisions by also touting nutritional or health qualities or both (29,32). Food advertising, like advertising for nearly all products, is regulated by the Federal Trade Commission (FTC). Historically, the FTC permitted nutrient claims (e.g., "high in fiber") in advertising and never formally prohibited diet-disease health claims (i.e., claims that explicitly linked the consumption [or lack of consumption] of a particular nutrient or other substance in a food to a disease or health-related condition [e.g., "a calcium-rich diet can help prevent osteoporosis"]) (32) . However, if diet-disease health claims were made on the label, the Food and Drug Administration (FDA) reclassified the food as a drug and required the manufacturer to adhere to the drugapproval procedures of the FDA (29). For years food advertisers did not make diet-disease health claims about their products, but as the connection between diet and health became increasingly clear, food manufacturers and advertisers grew interested in using this information to sell their products. Consequently, in 1984, the Kellogg Company initiated an advertising campaign that explicitly described the relationship between a high-fiber diet and reduced risk of certain types of cancer. When the FDA failed to prosecute this direct violation of 37 diet-disease health claims, other food manufacturers launched similar campaigns (18,29). Marketing strategies that included diet-disease health claims did provide consumers with information about nutrition and health. However, in their zeal to gain a competitive edge, advertisers also pushed the limits of what science could support and what consumers would believe (18,25). To stem questionable marketing practices and restore consumer confidence, the Nutrition Labeling and Education Act (NLEA) was passed in 1990 and became fully effective in 1994 (27). The NLEA overhauled nutrition labels on food packages, expanded the scope of nutrition labeling, explicitly defined nutrient content claims, and regulated dietdisease health claims (25). While the new food-labeling regulations did much to improve the quality of information on food packages, these regulations did not extend to food advertising (41). Fortunately, in its efforts to prevent deceptive or misleading claims, the FTC announced in 1994 that it would apply the standards set forth in the NLEA to evaluate nutrient content and diet-disease health claims made in food advertisements (14). The FTC reported that its goal was to create a food advertising policy that would help ensure that food advertising messages are congruent with data presented and are permitted on food labels (1 5). While food and beverage advertisements appear in all types of print, broadcast, and electronic media, television is the preferred advertising medium of food manufacturers--over 7 5 percent of their 1997 advertising budget was spent on televised advertising (17). The food and alcohol industry accounted for more than one-sixth of the $73-billion mass media advertising market; only the automobile industry spent more on advertising ( 17). 38 Although some studies have examined the nutrient content claims and health claims in food advertising, few have focused on televised advertising. Furthermore, no studies could be located that compared changes in nutrient content claims and health claims over time or examined the effect of the NLEA and FTC food advertising policy on televised food advertisements. Thus the purpose of this study was to examine changes in the nutrient content claims and health claims made in televised food advertisements before and after the enactment of the new food advertising policy of the FTC, a policy which is based on the NLEA, and to determine whether the use of claims varied by type of food product advertised. Methods Sample In the autumn ofboth 1992 and 1998, 17.5 hours of top-ranked, prime-time 1 were videotaped. This study focused on prime-time and major networks because they traditionally have the largest viewing audience (35). The sample comprised all commercials broadcast during the sampling period. Commercials (i.e., all non-program time) included advertisements, public service announcements, and promotions for television programs. Although all commercials were recorded and analyzed, only data pertaining to food advertisements are presented here. A food advertisement was defined as a paid-commercial announc;ement that specifically promoted a food, beverage, or dietary supplement intended for human consumption. 1 Prime-time refers to programming broadcast from 8 p.m. to 11 p.m. Monday through Saturday, and 7 p.m. to II p.m. on Sunday. Major networks refer to ABC, CBS, NBC, Fox, and WB; note WB became a network in 1998. Instrument The food advertisements were content and textually analyzed by using the study instrument that was adapted from those reported elsewhere (5,19,28,38, 40,50). Content analysis permits systematic, objective evaluation of visual and linguistic elements (6,24) . Textual analysis allows researchers to investigate how linguistic elements are used, their significance, and their contribution to understanding a topical area (4,38). Content analysis began by eliminating all nonfood commercials. All food advertisements were then classified into 11 food categories based largely on the USDA Food Guide Pyramid (47): Breads and cereals, vegetables, fruits, protein-rich foods (i.e., eggs, meat, poultry, fish, shellfish, nuts, and seeds), dairy products, high-sugar foods (e.g., syrup, candy, and soft drinks), high-fat foods (e.g., butter, oils, and salad dressing), alcohol-containing beverages (i.e., wine), calorie-free beverages, dietary supplements, and miscellaneous items (i.e., seasonings). Restaurant advertisements frequently highlighted a variety of food items that together comprised a meal. Thus to evaluate the nutritional value of the foods advertised, we assigned all items in an advertised meal to the appropriate food categories. In addition, combination foods (e.g., fast-food sandwiches and soups) were broken down into their component parts and appropriately assigned to two or more of the food categories. Foods in the first five categories listed previously were further classified by nutrient density: low, moderate, and high. Methods described in detail elsewhere were used to classify density {51). In brief, foods low in nutrient density tended to be ones that are highest in fat in each of the first five categories (e.g., pastries, French fries, coconut, luncheon meats, and whole milk). Foods moderate in Family Economics and Nutrition Review nutrient density were less nutrient dense than were foods high in nutrient density (e.g., breads made with enriched flour instead of whole grains, candied sweet potatoes instead of plain vegetables, fruits canned in syrup rather than fresh or canned in unsweetened juice, fattrimmed beef instead of skinless poultry white meat, or lowfat instead of nonfat milk). Foods high in nutrient density provided the greatest level of nutrients per kilocalorie. The subsequent step, requiring textual analysis, involved identifying and coding nutrient content claims as either (a) Contains Specific Nutrient or (b) Minimizes (or eliminates) Specific Nutrient. Nutrient content claims, defined in the FDA and USDA's foodlabeling regulations, include 11 core terms that can be used to describe the nutrient content of foods: good source, more, high, free, low, lean, extra lean, reduced, less, light, and fewer (42). An advertisement that indicated
Click tabs to swap between content that is broken into logical sections.
Title | Family Economics and Nutrition Review [Volume 13, Number 1] |
Date | 2001 |
Contributors (group) | Center for Nutrition Policy and Promotion (U.S.) |
Subject headings |
Home economics--United States--Periodicals Nutrition policy--United State--Periodicals |
Type | Text |
Format | Pamphlets |
Physical description | v. : $b ill. ; $c 28 cm. |
Publisher | Washington, D.C. : U.S. Dept. of Agriculture |
Language | en |
Contributing institution | Martha Blakeney Hodges Special Collections and University Archives, UNCG University Libraries |
Source collection | Government Documents Collection (UNCG University Libraries) |
Rights statement | http://rightsstatements.org/vocab/NoC-US/1.0/ |
Additional rights information | NO COPYRIGHT - UNITED STATES. This item has been determined to be free of copyright restrictions in the United States. The user is responsible for determining actual copyright status for any reuse of the material. |
SUDOC number | A 77.245:13/1 |
Digital publisher | The University of North Carolina at Greensboro, University Libraries, PO Box 26170, Greensboro NC 27402-6170, 336.334.5482 |
Full-text | CENTER FOR NUTRITION POLICY AND PROMOTION Ann M. V eneman, Secretary U.S. Department of Agriculture Eric M. Bost, Under Secretary Food, Nutrition, and Consumer Services Steven N. Christensen, Acting Deputy Director Center for Nutrition Policy and Promotion P. Peter Basi otis, Director Nutrition Policy and Analysis Staff The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, gender, religion, age, disability, political beliefs, sexual orientation, or marital or family status. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communication of program information (Braille, large print, audiotape, etc.) should contact USDA's TARGET Center at (202) 720-2600 (voice and TDD). To file a complaint of discrimination, write USDA, Director, Office of Civil Rights, Room 326-W, Whitten Building, 14th and Independence Avenue, SW, Washington, DC 20250- 9410 or call (202) 720-5964 (voice and TDD). USDA is an equal opportunity provider and employer. Editor Julia M. Dinkins Assistant Editor David M. Herring Features Editor Mark Lino Managing Editor Jane W. Fleming Contributor Joan C. Courtless Family Economics and Nutrition Review is writlen and published each quarter by the Center far Nutrition Policy and Promotion, U.S. Department of Agriculture, Washington, DC. The Secretary of Agriculture has determined that publication of this periodical is necessary in the transaction of the public business required by law of the Department. This publication is nat copyrighted . Contents may be reprinted without permission, but credit to Family Economics and Nutrition Review would be appreciated. Use of commercial or trade names does nat imply approval or constitute endorsement by USDA Family Economics and Nutrition Review is indexed in the following databases: AGRICOLA, Ageline, Economic Literature Index, ERIC. Family Studies, PAIS, and Sociological Abstracts. Family Economics and Nutrition Review is for sole by the Superintendent of Documents. Subscription price is $19.00 per year ($23.75 for foreign addresses). Send subscription order and change of address to Superintendent of Documents, P.O. Box 371954, Pitlsburgh, PA 15250-7954. (See subscription form on p. 127.) Original manuscripts are accepted far publication. (See "guidelines for authors" an back inside cover.) Suggestions or comments concerning this publication should be addressed to Julia M. Dinkins, Editor, Family Economics and Nutrition Review, Center far Nutrition Policy and Promotion, USDA. 31 01 Park Center Drive, 1Oth Floor, Alexandria, VA 22302. The Family Economics and Nutrition Review is now available at http:/ /www.cnpp.usda.gov. (See p. 126.) Research Articles 3 Profiles of Selected Target Audiences: Promoting the Dietary Guidelines for Americans Kay Loughrey, P. Peter Basiotis, Claire Zizza, and Julia M Dinkins 15 U.S. Teens and the Nutrient Contribution and Differences of Their Selected Meal Patterns Anna Maria Siega-Riz, Claude Cavadini, and Barry M. Popkin 27 Associations Between the Milk Mothers Drink and the Milk Consumed by Their School-Aged Children Rachel K. Johnson, Celeste V. Panely, and Min Qi Wang 37 The Effects of Food Advertising Policy on Televised Nutrient Content Claims and Health Claims Carol Byrd-Bredbenner and Darlene Grasso 50 The Thrifty Food Plan, 1999: Revisions of the Market Baskets Staff at the Center for Nutrition Policy and Promotion-Compiled by Mark Lino 65 Sample Menus and Recipes Based on the 1999 Thrifty Food Plan Staff at the Center for Nutrition Policy and Promotion- Compiled by Myrtle Hogbin and Mark Lino Research Briefs 77 USDA's Food Guide: Updating the Research Base to Reflect Changes in Food Consumption Patterns Kristin L. Marcoe 81 USDA's Expenditures on Children by Families Project: Uses and Changes Over Time Mark Lino 87 Current Knowledge of the Health Effects of Sugar Intake Anne L. Mardis 92 Insight 11 : Food Portions and Servings: How Do They Differ? Myrtle Hogbin, Anne Shaw, and Rajen S. Anand 95 Insight 14: A Focus on Nutrition for the Elderly: It's Time to Take a Closer Look Nancy W. Gaston, Anne Mardis, Shirley Gerrior, Nadine Sahyoun, and Rajen S. Anand 98 Insight 19: Beliefs and Attitudes of Americans Toward Their Diet Julia M. Dinkins Volume 13, Number 1 2001 Research Summaries 101 Consumer Price Index Research Series Using Current Methods 1 05 Changes in the Health Services Industry 108 Extended Measures of Well-Being: Meeting Basic Needs 112 Measuring Time at Work Regular Items 1 15 Research and Evaluation Activities in USDA 118 Federal Studies: Review of the Nutritional Status of WIC Participants 120 Journal Abstracts 122 Official USDA Food Plans: Cost of Food at Home at Four Levels, U.S. Average, May 2001 123 Official USDA Alaska and Hawaii Thrifty Food Plans: Cost of Food at Home (2nd Half 2000) 124 Consumer Prices 125 U.S. Poverty Thresholds and Related Statistics Kay Loughrey, MPH, RD MRP P. Peter Basiotis, PhD Center for Nutrition Policy and Promotion Claire Zizzo, MS University of North Carolina at Chapel Hill Julia M. Dinkins, PhD Center for Nutrition Policy and Promotion 2001 Vol. 13 No. I Research Articles Profiles of Selected Target Audiences: Promoting the Dietary Guidelines for Americans To decrease the risk of nutrition-related diseases, Americans need to narrow the gap between scientifically based nutrition guidance and their nutrition-related behaviors. This study examines the usefulness of segmentation and audience-profiling techniques in promoting the Dietary Guidelines, designed to help narrow this gap. Using the 1991- 94 survey of the Market Research Corporation of America Information Services (MRCA), we segmented 491 women gatekeepers into tertiles (Better Eaters, Fair Eaters, and Poor Eaters) based on their scores on a modified version of the Healthy Eating Index. We then compared the segments' demographic characteristics; health and diet orientation; values about, and perceived benefits and barriers to, healthful eating; nutrition, food preparation, and shopping habits; and media habits. Results showed that women gatekeepers were interested in improving their diets, and they differed significantly regarding values, benefits, and barriers of eating a healthful diet and nutrition; food preparation practices; and shopping habits. We discussed the implications of these differences in terms of improving the quality of the diet. T he Dietary Guidelines for Americans (23), issued by the U.S. Departments of Agriculture (USDA) and Health and Human Services (DHHS), answer this basic question: "What should people eat to stay healthy?" Forming the basis of Federal nutrition policy affecting food, nutrition education, and information programs, the Guidelines stress the significance of dietary balance, variety, and moderation (7). Still, in the United States, four of the leading causes of death-heart disease, cancer, stroke, and diabetes-are linked to nutrition (1 0). Americans still need to increase total intake of fruits, vegetables, and grain products and to decrease intake of fat and saturated fat. Although some progress has been made based on progress in meeting Year 2000 Objectives, the startling increase in the portion of Americans who are overweight or obese poses one of the biggest challenges in meeting Healthy People 2000 (24). A summary measure of dietary status- the Healthy Eating Index-has shown that 7 of I 0 Americans need to improve their diet (4) . Other results have also indicated that although Americans choose a wide variety of foods, they consume less than the recommended servings from the fruit, dairy, meat, grains, and vegetable groups of the Food Guide Pyramid. Americans' consumption of calories from fats and sugars, however, exceeds Pyramid recommendations (11). 3 Thus evidence has shown that Americans still need to improve their diet; Americans need to narrow the gap between scientifically based nutrition guidance and consumer behavior that may increase the risk of illness from nutrition-related diseases. To better meet the needs of the public, some authors believe the Guidelines need to do two things: (1) continue to advance national dietary guidance that is based upon scientific evidence and (2) promote dietary guidance in ways that will lead to behavior change, improved health, and nutritional well-being (22) . The purpose of this study is to examine the extent to which major differences exist between audience segments on key variables, to profile these audience segments, and to suggest whether these differences warrant distinct nutrition education approaches in attempting to change dietary behaviors. We describe three segments of female gatekeepers and how their characteristics differ on several dimensions: demographic and health status; values about, and benefits and barriers to, healthful eating; nutrition, food preparation, and shopping habits; and media habits. We discuss the implications of these differences in terms of improving the dietary behavior of these segments. We believe that nutrition educators can directly apply this information when they design program interventions. The underlying assumption of social marketing and marketing approaches is that different audience segments require alternate approaches for achieving a desired behavior change. This study examines whether this assumption applies to nutrition education to create dietary behavior change. 4 Lastly, we examined results in relationship to behavioral models and theories. We examined how the segments might differ with respect to their stage of behavior change and the extent to which audience segments could be described, based on Prochaska and Di Clemente's trans theoretical model of change (18). The stages of this model are precontemplation (not considering whether to make a change), contemplation (thinking about making a change), decision (making definite plans to change), action (initiating change), and maintenance. This model has been used to describe dietary behavior in relationship to weight control and the reduction of dietary fat (6, 19). We looked to social learning theory, which is based on social cognitive theory, to inform recommendations for designing strategies for behavior change (3,17). Social learning theory emphasizes the interaction of cognition, other personal factors (e.g., selfefficacy), and environmental factors on behavior. Several critical personal factors suggested by social learning theory have been assessed in this analysis: • Perception of the situation • Anticipated outcomes of behavior • Knowledge and skills to perform a behavior • Confidence in performing a behavior We considered the theory of planned behavior in forming program implications (1). This theory suggests that people will be more likely to take action if it leads to consequences they desire. It also suggests that behavior and behavioral intent are influenced by the degree of control people think they have over circumstances and their ability to perform a behavior. Background Research indicates that nutrition promotion of the Guidelines should focus on behavior change; have a strong consumer orientation; segment and target consumers; use multiple, reinforcing, interactive channels; and refine consumer messages continually (22,23). Segmentation, a frequently used approach in commercial-sector marketing, has been used in programs designed to change health behaviors (2) and has been used to create a profile or snapshot that represents the target audience. It, as well, has encouraged creative communication that is tailored to the target audience (6,12,15). To segment audiences, social marketers analyze potential markets and create subgroups of target populations with similar characteristics regarding the desired behavior. Then they allocate resources among one or more subgroups and vary the methods used to reach each subset (2). Health communicators also use segmentation methods to identify people who are similar in key respects and to tailor the content and delivery of the communication based on people's profiles (16,21). Target-audience profiles have been used in large-scale nutrition education programs, including the 5 A Day media campaign of the National Cancer Institute (1 3,15) and the Nutrition and Physical Activity program of the Centers for Disease Control (9). Family Economics and Nutrition Review Methods Database We analyzed data from the 1991-94 survey of the Market Research Corporation of America Information Services (MRCA). Nationally representative, the MRCA survey consists of information on people's food and beverage consumption and their opinions and attitudes about general interests, health, diet and food preparation, shopping, and media usage. The MRCA data set consists of five surveys and two database systems: Household Information Form, Menu Census Diaries, Pyschographic Questionnaire, Diet Information Quiz, and Food and Nutrition Attitude Inventory. To select participating households, MRCA uses a multistage, stratifiedrandom procedure. In stage !-the Household Recruiting Pool-a sampling pool of households is generated from generic consumer listings of U.S. households of various demographic types. Households that agree to participate then qualify for the second stage of sampling-the National Consumer Panel. The Panel consists of 5,000 households whose demographic characteristics (household size, homemaker age, household income, census regions, and metro-area size) are matched to the U.S. Census. The third stage-the Menu Census Panelconsists of a subsample of households (n=2,000) from the National Consumer Panel. For the Menu Census Panel, MRCA uses a stratified-random procedure to select 500 households each quarter. Detailed food diaries of food and beverage consumption are collected for 14 consecutive days. Actual serving sizes are not collected. They are imputed based on eating occasions for individual foods by applying standard serving sizes. For this reason, they should be considered 2001 Vol. 13 No. 1 estimates rather than precise measures of food and beverage consumption. The Nutrient Intake database measures macro- and micro-nutrient intake; the Food Guide Pyramid database measures "servings"' of the Pyramid Food Groups. Healthful Eating Measure The USDA Healthy Eating Index (HEI) measures the overall quality of Americans' diet (4) and uses data from the USDA Continuing Survey ofFood Intakes by Individuals (CSFII). The HEI uses 1 0 components to measure different aspects of a healthful diet: • Components 1-5 measure the degree to which a person's diet conforms to serving recommendations of the food groups of the USDA Food Guide Pyramid: Grains (bread, cereal, rice, and pasta), vegetables, fruits, milk (milk, yogurt, and cheese), and meat (meat, poultry, fish, dry beans, eggs, and nuts). • Components 6 and 7 measure consumption of total fat and saturated fat, respectively, as a percentage of total food energy intake. • Component 8 measures total cholesterol intake. • Component 9 measures sodium intake. • Component 10 measures the variety of a person's diet on any given day. 1 MRCA used total frequency of"eatings" as the main measure of the individual food consumed. MRCA estimated serving sizes for each eating occasion for over 330 collapsed food categories based on 1987-88 USDA data on number of grams for each eating occasion for individual food items. MRCA then assigned different serving sizes to 18 age-gender groups: four age groups for children under 12 and seven age groups each for males and females over age 13. Americans still need to improve their diet; Americans need to narrow the gap between scientifically based nutrition guidance and consumer behavior that may increase the risk of illness from nutrition-related diseases. 5 Each component of the HEI has a maximum score of 10 and a minimum score of zero; intermediate scores are computed proportionately. The maximum overall score for the 10 components combined is 100. Higher component scores indicate intakes close to recommended ranges or amounts. The MRCA does not provide information on variety; hence, we used a modified version of the HEI to examine characteristics that distinguish women from the MRCA sample with higher quality diets from those with lower quality diets. All scores on the modified version were adjusted to a 1 00-point score. Thus the total maximum score was 100. To compute individual HEI scores, we matched the female gatekeeper to the appropriate serving recommendations of the Pyramid Food Groups. We calculated gatekeepers' average percentage of calories from total fat and saturated fat and compared their intakes of cholesterol and sodium with Pyramid recommendations. Sample We selected healthy adult women in the United States as the unit of analysis (target audience) because they often are gatekeepers who shape their family's nutrition and health habits. Our sample consisted of women gatekeepers aged 25 through 55, reporting household income of $20,000 to $125,000 and no major health problems. Those excluded reported having high blood pressure, diabetes, heart disease, high levels of serum cholesterol, or followed a diet for diabetes or allergies. We could not use marital status as a screening variable because MRCA does not include information on respondents' marital status. The database also does not include information on vegetarian diets, employment status or profession, 6 and the relationship of household members. The final sample of 491 gatekeepers was weighted to reflect the U.S. population of interest. After ranking and dividing the gatekeepers into tertiles (segments) based on their scores on the modified HEI, we developed profiles of the women gatekeepers and used multiple t tests to examine differences among the three segments. SUDAAN (Software for the Statistical Analysis of Correlational Data), which accounts for sampling designs that are complex and stratified, was used in the analysis to ensure appropriate estimates of standard errors for hypotheses testing.2 Results Demographic Characteristics The women gatekeepers who were Better Eaters (having the highest HEI score) are the basis of comparison with other groups of women gatekeepers: Fair Eaters and Poor Eaters. The women gatekeepers differed in some ways (table 1). Compared with the other groups, the Better Eaters more closely met the recommendations of the USDA Food Guide Pyramid. Based on percentages, overall, the women gatekeepers' average Healthy Eating Index score was 57 percent. With an average score of74 percent, the Better Eater had the higher HEI score, followed by the Fair Eater, with 62 percent; and Poor Eater, with 52 percent. Healthy Eating Index scores were calculated based on the degree to which a person in the sample's diet 2 "SUDAAN is specifically designed for analysis of cluster-correlated data from studies involving recurrent events, longitudinal data, repeated measures, multivariate outcomes, multistage sample designs, stratified designs, unequally weighted data, and without replacement samples" (20). conformed to serving recommendations of the food groups of the USDA Food Guide Pyramid as previously described. There are small differences in the gatekeepers' average years of education, height, Body Mass Index (BMI), likelihood of having children present in the household, and race. The Better Eater was more likely than the other Eaters to have more years of education. Compared with the Poor Eater, the Better Eater had a lower BMI score, was slightly taller, and more likely to be White or of a race other than Black. Compared with the Fair Eater, the Better Eater was less likely to have children. The women gatekeepers had some characteristics in common (tables 1 and 2). Their characteristics were considered similar if more than 60 percent of the women in each group exhibited them and if the differences in the characteristics were statistically insignificant (p>.Ol). These three groups were similar demographically based on age, household size, household income, and self-reported weight. Values, Benefits, and Barriers to Healthful Eating Similar to the Better Eater, the Fair Eater (F) reported that eating a healthful diet was important to her (table 3). Both said they could avoid future health problems-a perceived longterm benefit-by eating more healthfully. Similarly, the Fair Eater and the Better Eater reported that eating "healthy foods" gave them the energy they needed- a perceived shortterm benefit- and agreed that eating "healthy foods" improved their physical appearance. Family Economics and Nutrition Review Table 1 . Education distinguishes all three segments of women gatekeepers: Demographic and health status variables, MRCA 1991-94 Diet status Variable Better Eaters Fair Eaters Poor Eaters Mean Age (years) 39 38 38 Household size 3.3 3.3 3.3 Household income (thousands) 42.97 41.40 41.93 Education (years) 14.2* 13.7* 13.2* Weight (kg) 67.39 67.96 71 .65 Height (em) 164.7* 163.9 162.7* BMI 25.07* 25.54 27.31* Percent HEI score' 74 62 52 Children present 56* 72* 65 White 94.9* 87.5 83.6* Block 4.1 7.3 9.3 Other 1.0* 5.2 7.1* 'The Healthy Eating Index scores differ because this foetor was used to segment the women gatekeepers. *Means or percentages within the some row ore significantly different (p < 0.05). The Fair Eater differed, however, from the Better Eater in two important ways. ( 1) She was Jess likely than the Better Eater to believe she could avoid future health problems by exercising. (2) Both convenience and taste were barriers for the Fair Eater, who was more likely than the Better Eater to say that "healthy foods" had to be convenient for her to use them and to report that a reason for not choosing healthful foods was because they didn't taste good. The Poor Eater (P) was less likely than the Better Eater to believe it was important to eat a healthful diet, look and feel physically fit, maintain a proper weight, and to identify with potential benefits of healthful eating. She was less likely to agree that she could avoid future health problems by eating a healthful diet and by exercising; she was less likely to report the perceived short-term benefit that eating "healthy foods" gave her the energy she needed and improved her physical appearance. The Poor Eater also 2001 Vol. 13 No. 1 indicated that she was less likely than her counterpart to say she knew how to eat healthfully. She was, however, more likely than the Better Eater to report that eating healthfully was too complicated and confusing. Health and Diet Orientation All of the women gatekeepers believed they were knowledgeable about health and nutrition (table 2). They reported an interest in improving their diets, agreed they had some weight to lose, and tried to do so, at least occasionally. Similarly, they agreed that it was important for them to live long and healthy lives. Nutrition, Food Preparation, and ·Shopping Habits Similar practices among the women gatekeepers extended to how they shopped for food and planned and prepared it (table 2). Among the many similarities, all three groups redeemed the coupons they clipped from magazines and newspapers. The Poor Eater was less likely than the Better Eater to believe it was important to eat a healthful diet, look and feel physically fit, maintain a proper weight, and to identify with potential benefits of healthful eating. 7 Table 2. Better Eaters, Fair Eaters, and Poor Eaters have many characteristics' in common, MRCA 1991-94 Variable Health and diet orientation Physical activity Pyschographics Shopping Food planning and preparation Family eating habits Media Commonalities Believe they ore knowledgeable about health and nutrition Interested in improving their diets Think they hove some weight to lose Try, at least occasionally, to lose weight Believe it is important for them to live a long, healthy life Frequency Like to meet new people Join actively in community groups Desire to be well respected Like the outdoors Enjoy taking the family to a different vocation spot each year Make a complete list before going shopping Enjoy browsing through supermarket aisles Do not like the excitement of a busy supermarket Save a lot of money by shopping around for food bargains Stock up on named brand foods that they like during soles Cut coupons out of newspapers and magazines Redeem coupons (almost always) Send away for items offered through advertising Willing to pay for certain food items for special occasions En joy cooking and think of themselves OS creative cooks Don't like to bother cooking just for themselves (when alone) Enjoy preparing a fancy meal for their families once in awhile Collect recipes from the food sections of the newspapers Exchange recipes with friends and relatives Add something extra (almost always) to prepared foods Serve the some evening meals from one week to the next Try to make use of leftovers but usually throw them out Hove some family members who ore concerned about being overweight View television-network evening news, cable news/ television Read magazines and newspaper 'Characteristics were common if more than 60 percent of each group exhibited them and if the differences in the characteristics were statistically insignificant (p > .01 ). 8 Family Economics and Nutrition Review Table 3. Most measured beliefs and practices of Poor and Fair Eaters differ from those of Better Eaters, MRCA 1991-94 Degree to which Poor (P) and Fair (F) Eaters soy the following, compared with Better Eaters Variable As likely More likely Less likely Values, Benefits, and Batriers Eating a healthy diet is important to me. F I can ovoid future health problems by eating healthfully. F I choose healthy foods because they give me the energy I need. F I choose healthy foods because they improve my physical appearance. F Healthy foods hove to be convenient for me to use them . A reason for not choosing healthy foods is they don't taste good. Trying to eat healthy is too complicated and confusing . I con ovoid future health problems by exercising. It is important for me to look and feel physically fit. It is important for me to maintain my proper weight. I know how to eat healthy. Nutrition, Food Preparation, and Shopping Habits I worry about the nutritional content of the foods I eat. F I always see to it that my family takes vitamins. P I'm much more willing to try a new recipe when someone I know tried it and liked it. I always or usually pay attention to on-shelf, aisle display. Most snack foods I like ore unhealthy. I do not discuss various foods and their food values with my family so they understand nutrition better. I always pay attention to instant coupons. I make every possible effort to see that my family eats really nourishing foods. I get upset if the family doesn't eat together. I go out of my way to buy non-fat foods. Frozen foods ore more nutritious than conned foods. I serve fish because it has less fat. I disagree that red meat is better for your health than fish . I do not look for prepared dishes when I shop. I collect recipes from magazines. I disagree that my family is easy to please. Media I watch television in general, including entertainment programs, and daytime television . I watch television programs like police/ private eye and daytime serials because I really like them . I watch television serials/soap operas because I like them . I watch prime-time television programs. I read women's general interest magazines. F F F, p F F p p P, F F, p F, p F P, F p p p F, p p p p p F F, p F, p F, p F, p F, p F p p p p Note: The "F" and "P" for the Fair Eaters and Poor Eaters, respectively, indicate that these women gatekeepers differ significantly from the comparison group: the Better Eaters, at the 0.01 level. 2001 Vol. 13 No. 1 9 Compared with the Better Eater, the Poor Eater was less likely to worry ~bout the nutritional content of the foods she ate. 10 The groups differed, however, in a number of important ways related to nutrition, food preparation, and shopping habits (table 3). Similar to the Better Eater, the Fair Eater worried about the nutritional content of the foods she ate. Still, she was less likely than the Better Eater to make an effort to serve her family nourishing foods, get upset if the family didn't eat together, and go out of her way to buy nonfat foods. She was more likely than the Better Eater to pay attention to on-shelf, aisle display ads and instant coupons and to look for prepared foods when shopping. Compared with the Better Eater, the Poor Eater was less likely to worry about the nutritional content of the foods she ate. Like the Fair Eater, she was also less likely than the Better Eater to make every possible effort to see that her family ate nourishing foods, to get upset if the family didn't eat together, and to go out of her way to buy nonfat foods. The Poor Eater was more likely than the Better Eater to pay attention to instant coupons, to agree that most of the snack foods she liked were unhealthful, and to disagree that she discussed foods with her family so they understood nutrition better. Media The three groups watched similar television programs or stationsevening network news, cable news, and cable TV-and they read similar magazines and newspapers (table 2). However, the Fair Eater and Poqr Eater were more likely than the Better Eater ~o watch television in general, includmg entertainment (non-news) shows and daytime programs (table 3). The Poor Eater also watched less primetime television than did the Better Eater and was less likely to read women's general interest magazines. Discussion Profiles Demographic differences in audience segments do not explain the overall differences in the three segments' approaches to food consumption. Results of this analysis indicate a small number of demographic differences. Then what might explain these differences? The Better Eaters are more likely than the Poor Eaters to report that eating a healthful diet is important to them and are concerned about the nutritional content of their diets. They are likely to perceive short- and long-term benefits of eating healthfully, and are taking action to eat healthfully. Better Eaters are categorized in this analysis as being either in the action or maintenance stages of the transtheoretical model of change, though direct assessment of the stages of change was not measured in this analysis. Better Eaters are considered in one of these two stages of change based on their HEI score, their concerns about nutrition, and their greater tendency to act on their concerns. It is therefore not possible to determine precisely whether they are in the action or maintenance stage, using the algorithm applied by Curry et al. for staging dietary fat reduction (6). In terms of social learning theory, Better Eaters appear to be able to anticipate the outcomes of their behavior and self-determine their behavior, successfully although not perfectly. They appear to be confident of their ability to carry out healthful eating behaviors based on their being less likely to report that trying to eat more healthfully is complicated and confusing than did women in the other two segments. Better Eaters experience Family Economics and Nutrition Review a rather high degree of control over their circumstances in terms of eating healthfully, based on their responses to all questions, collectively. This characteristic is a key factor in the theory of planned behavior. Still, Better Eaters have room for improving their diets based on their HEI scores. Fair Eaters, compared with Better Eaters, report a mixture of benefits, barriers, and actions that may account for their lower HEI score. Like the Better Eaters, Fair Eaters are more likely than the Poor Eaters to report that eating a healthful diet is important to them, and are concerned about the nutritional content of their diets. They are as likely as Better Eaters to perceive short- and long-tenn benefits of eating healthfully, and are taking some action to eat healthfully. However, they are less likely to go out of their way to eat healthfully, such as making an effort to serve their families nourishing foods and buying nonfat foods. They are more likely to respond to in-store promotions such as on-shelf, aisle display ads, and instant coupons. Taste and convenience are especially important to Fair Eaters, and they are more likely than Better Eaters to select prepared foods. In terms of media use, they are more likely to watch television, particularly for entertainment. Lastly, the Fair Eaters are more likely to report that eating healthfully is complicated and confusing, compared with Better Eaters. In sum: Fair-Eaters are convinced yet not committed to eating healthfully. While they are interested in the positive results associated with eating healthfully and are convinced of its benefits, Fair Eaters are less proactive in making healthful eating choices, and appear to respond passively to stimuli in their environment, be it family, in-store cues, desire for sensory satisfaction, or ease in meal preparation. As a group, they 2001 Vol. 13 No. I appear to eat healthfully when it's convenient and could be characterized as "convinced, but not committed" to eating healthfully. Many factors can intervene in their environment to prevent them from eating healthfully. Fair Eaters could be considered to be in a late stage of contemplation in terms of stages of change, although screening questions for staging were not included in the original MRCA questionnaire. No questions were asked that could help determine whether Fair Eaters had developed a plan of action that would place them in the preparation stage of the transtheoretical model of change. Still, their passivity in relationship to environmental cues indicates that they have not developed a concerted plan of action that they intend to implement in the near future. In terms of social learning theory, Fair Eaters are aware of the outcomes of behaviors, including expected results and benefits but lack the knowledge and confidence to eat more healthfully based on the fact that, compared with Better Eaters, they are more likely to report that trying to eat healthfully is too complicated and confusing. They also seem to experience a rather low degree of control over their circumstances, an important factor influencing their behavior that is emphasized by the theory of planned behavior. A number of factors may prevent Poor Eaters from taking actions that could improve their dietary habits, factors that may account for their HEI scores being the lowest among these three groups. They are less likely to report an interest in achieving results related to healthful eating. For example, they are less likely to report that eating more healthfully is important to them, compared with Better Eaters. Poor Eaters are also less likely to be convinced of long-term benefits: they are less likely than Better Eaters to agree that they can avoid future health problems by eating a healthful diet. Nor are they convinced of short-term benefits such as being less likely to agree that "healthy foods" give them the energy they need. They are also less likely to know how to eat healthfully and are more likely to perceive that eating healthfully is complicated and confusing. Poor Eaters are less concerned about nutrition for themselves and their families: they are less likely to report that they worry about the nutrient content of the food they eat. They are also less likely to talk with their families about foods in terms of their nutritional value or to report making every possible effort to see that their families eat nourishing food. Thus Poor Eaters are somewhat interested in improving their diets, but are not convinced of the benefits of doing so. They are also less concerned with achieving the potential results of eating healthfully than are Better Eaters. While they, like other gatekeepers, claim to be knowledgeable about health and nutrition, they admit to not knowing how to eat healthfully. They could be characterized as "interested but unconvinced" that healthful eating is particularly relevant to them. Poor Eaters could be categorized as being in an early phase of contemplation (transtheoretical model of change) based on their interest in improving their diet. Although Poor Eaters appear to be aware of where they stand when it comes to eating healthfully, they lack three key critical personal factors described by social learning theory: (1) the ability to anticipate outcomes of their behavior, (2) knowledge and skills to act, and (3) confidence to perform this behavior. 11 Program Implications Given the large number of characteristics these three segments of women have in common, should the same approach to nutrition education be used for these three groups? Speaking in favor of a common approach are the characteristics the three segments share. However, many of the characteristics the three audience segments have in common may be attributed to the fact that the segments are all primary food preparers. A number of important differences among these three segments of women discussed in this paper suggest that different approaches to nutrition education are needed for each segment. For the Better Eaters especially, providing tips that are simple, positive, and easy to apply may build on their current interest and actions to improve their diets. A different approach should be used with Fair Eaters. Nutrition education for this group should appeal to their interest in taste and convenience. Communication and education strategies should be used to deliver actionable messages and illustrate easy methods for improving their diet that do not sacrifice taste. Suggestions should be offered that are easy to apply such as adding a grated carrot to prepared tomato sauce as a way to add sweetness, improve its taste, and add important nutrients. It may also be helpful to highlight convenient ways to more healthful eating such as offering ideas that they can do quickly such as a "10-minutea- day" way to improving their eating habits. Fair Eaters should be targeted with a few carefully selected nutrition messages that are easy to understand and apply, and that are likely to cut through confusion generated by media coverage of nutrition news. Nutrition education for Fair Eaters should use mass media to remind them frequently about eating healthfully. It should also be presented in an entertaining way, 12 because this audience is used to regular television entertainment. It will require a highly targeted approach to reach Poor Eaters with nutrition education. An approach is needed that immediately captures their attention and establishes cultural and lifestyle relevance. To help establish relevance of consequences of healthful eating, messages to this audience should come from people they perceive as peers or from someone they admire, such as a celebrity, who can model the desired behavior. For example, the Milk Mustache Campaign has shown celebrities and opinion leaders with their milk mustaches as a way to establish that drinking milk is a highly acceptable and desirable behavior with their target market. Nutrition education programs and materials that are highly targeted to a specific lifestyle or cultural experience are likely to be welcomed. For example, the National Cancer Institute developed and tested Down Home Healthy, a recipe booklet designed for an African-American audience, and found that respondents were highly interested in this book because of its cultural relevance (8). Introducing this recipe booklet was used to explore interest in an approach of encouraging African Americans to use modified versions of traditional recipes to lower fat and increase fiber intake. Responses to the recipe booklet and accompanying brochure were the most active and engaging aspects of focus group sessions. Participants welcome9 this approach, if the taste of the food presented in the recipes met their expectations. Successful nutrition education strategies are recommended that will break abstract nutrition concepts into practical action steps that can easily be mastered and applied to help build knowledge, skills, and self-efficacy for eating more healthfully. For example, guidance about adding more fiber to the diet should include a brief discussion of the Nutrition Facts panel of the food label. It should include making a specific request to ask people to go to the grocery store and compare the fiber content on the food label of several breakfast cereals they like, and then purchase a cereal that contains 20 percent or more of the Daily Value for fiber per serving. This approach was highly effective in transforming apathy into keen interest in nutrition among working and middle-class women attending focus groups sponsored by the National Cancer Institute (1 4). This segment of women gatekeepers, in particular, may be encouraged to begin taking action as they experience more short-term benefits that are meaningful and motivating. To accomplish this, nutrition education and promotion efforts for Poor Eaters should move them from being interested to being convinced that healthful eating is meaningful and relevant to them. Summary The most effective ways to reach these women gatekeepers by segment is as follows: 1. Better Eaters: Offer new tips that can be added to their current actions for eating healthfully. 2. Fair Eaters: Insert frequent environmental cues to eating healthfully that will appeal to their interest in taste and convenience. 3. Poor Eaters: Establish relevance by identifying ways to appeal immediately to this audience that are consistent with their lifestyle and cultural context. Family Economics and Nutrition Review These fmdings are consistent with those of authors reviewing nutrition education for adults (5). In their review of successful nutrition education interventions for adults, the authors suggested nutrition education communication and strategies in programs that • Are ongoing and multifaceted; • Use mass media to increase awareness and enhance motivation; • Tailor strategies based on formative audience research; • Use motivational messages and educational strategies; and • Employ a behaviorally focused approach that is based on personal factors, behavioral capabilities, and environmental factors. The results of this study suggest that nutrition educators can apply the same segmentation methods used by social marketers and health communicators. It can be expected that doing so would allow them to make the most effective use of resources and to increase program efficiency. We suggest that with a greater understanding of applicable target segments, nutrition educators, policymakers, and other information multipliers will be betterpositioned to improve the diets of Americans. 2001 Vol. 13 No. 1 References I. Ajzen, I. and Madden, T.J. 1986. Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control. Journal of Experimental Social Psychology 22:453-4 72. 2. Andreasen, A.R. 1995. Marketing Social Change. Jossey-Bass Publishers, San Francisco, CA. 3. Bandura, A. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice-Hall, Englewood Cliffs, NJ. 4. Bowman, S., Lino, M., Gerrior, S., and Basiotis, P. 1998. The Healthy Eating Index 1994-96. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion. CNPP-5. 5. Contento, 1., Balch, G.I., Bronner, Y.L., Paige, D.M., Lytle, L.A., Maloney, S.K., Olson, C.M., and Swadener, S.S. 1995. Nutrition education for adults. Journal of Nutrition Education 27(6):312-328. 6. Curry, S.J., Krista!, A.R., and Bowen, D.J. 1992. An application of the stage model of behavior change to dietary fat reduction. Health Education Research 7:97-105. 7. Davis, C. and Saltos, E. 1996. The Dietary Guidelines for Americans- Past, present, future. Family Economics and Nutrition Review 9{2):4-13. 8. Eisner, E., Loughrey, K., and Hairston, B. 1995. Focus Groups with African Americans on Nutrition and Cancer. National Cancer Institute, Bethesda, MD. 9. Federal Program Update. 1996. Nutrition and physical activity communications at the Centers for Disease Control and Prevention. Social Marketing Quarterly 3: 14-15. 10. Frazao, E. 1996. The American diet: A costly health problem. FoodReview 19(1):2-6. II . How Do Americans ' Diets Measure Up To The 1995 Dietary Guidelines For Americans? July 1998 (7). U.S. Department of Agriculture, Food Surveys Research Group. 12. Kover, A. 1995. Copywriters' implicit theories of communication: An exploration. Journal of Consumer Research 21 :596-611. 13. Lefebvre, R.C., Doner, L., Johnston, C., Loughrey, K., Balch, G.I., and Sutton, S.M. 1995. Use of database marketing and consumer-based health communication in message design: An example from the Office of Cancer Communications "5 A Day for Better Health" program. In E. Maibach and R.L. Parrott (Eds.), Designing Health Messages: Approaches from Communication Theory and Public Health Practice (pp. 217-246). Sage, Thousand Oaks, CA. 14. Loughrey, K., Eisner, E., Doner, L., and Weinberg, L. 1994. Exploring the Nutrition Facts Food Label As a Too/for Educating Consumers about NCI's Dietary Guidelines, A Focus Group Study. National Cancer Institute. 15. Loughrey, K.A., Balch, G.I., Lefebvre, R.C., Doner, L., Johnston, C., Eisner, E., and Hadley, L. 1997. Bringing 5 A Day consumers into focus: The qualitative use of consumer research to guide strategic decision making. Journal of Nutrition Education 29(4): 172-177. 13 16. Maibach, E. W., Maxfield, A., Ladin, K., and Slate, M. 1996. Translating health psychology into effective health communication: The American Healthstyles Audience Segmentation Project. Journal of Health Psychology I :261-277. 17. Perry, C.L., Baranowski, T., and Parcel, G.S. 1990. How individuals, environments, and health behavior interact: Social learning theory. ln K. Glanz, F.M. Lewis, F., and B.K. Rimer (Eds.), Health Behavior and Health Education Theory, Research, and Practice (pp. 161-186). Jossey-Bass Publishers, San Francisco, CA. 18. Prochaska, J.O. and DiClemente, C.C. 1982. Transtheoretical therapy: Toward a more integrative model of change. Psychotherapy: Theory, Research and Practice 19:276-288. 19. Prochaska, J.O. and DiClemente, C. C. 1985. Common processes of self-change in smoking, weight control, and psychological distress. ln S. Shiffman and T. Wills (Eds.). Coping and Substance Use (pp. 345-363). Academic Press, Orlando, FL. 20. Research Triangle Institute. 1998 (July 20). Why SUDAAN? [On-line]. Available: http: //www. rti.org/patents/sudaanlwhy sudaan.html. 21 . Slater, M.D. 1995. Choosing audience segmentation strategies and methods for health communication. In E. Maibach and R.L. Parrott (Eds.). Designing Health Messages: Approaches From Communication Theory and Public Health Practice (pp. 186-198). Sage, Thousand Oaks, CA. 22. Sutton, S.M., Layden, W., and Haven, J. 1996. Dietary guidance and nutrition promotion: USDA's renewed vision of nutrition education. Family Economics and Nutrition Review 9(2): 14-21 . 23. U.S. Department of Agriculture and U.S. Department of Health and Human Services. 2000. Nutrition and Your Health: Dietary Guidelines for Americans (5th ed.). U.S. Department of Agriculture. Home and Garden Bulletin No. 232. 24. U.S. Department of Health and Human Services. 2000. Healthy People 2010. Washington, DC. 14 Family Economics and Nutrition Review Anna Maria Siega-Riz, PhD Department of Nutrition School of Public Health University of North Carolina at Chapel Hill, NC, USA Claude Cavadini Nestle Research Center Lausanne, Switzerland Barry M. Popkin, PhD Department of Nutrition School of Public Health University of North Carolina at Chapel Hill, NC, USA 2001 Vol. 13 No. 1 U.S. Teens and the Nutrient Contribution and Differences of Their Selected Meal Patterns We examined the nutrient contribution of foods consumed at breakfast, brunch, lunch, dinner, and snacks, as well as the types of foods consumed on those occasions, by adolescents (n= 1,31 0) participating in the 1989-91 Continuing Survey of Food Intakes by Individuals. Descriptive statistics were generated, using weights and taking into account sample design effects, to examine the consistency of their meal patterns: Consistent, moderately consistent, and inconsistent. Results showed that for individuals with an inconsistent meal pattern, dinner provided half of the day's energy and total snacks provided over one-fifth, an equivalent of one meal for others. Most nutrients studied (fat, protein, calcium, and iron) followed the same pattern as energy. Age differences were noted : 15- to 18-year-olds were more likely to have inconsistent patterns. The types of foods consumed also differed by meal pattern. Both increasing the consistency in the number of meals consumed, as well as improving food-selection behaviors, may serve as possible interventions to improve the diets of adolescents. A dolescence is a period of great transitions. Nutrient require ments are increased from childhood because of physical growth, and behaviors acquired during this period persist into adulthood (1,11, 17,22). While many subsets of adolescents engage in behaviors that have wide public health attention, some adolescents may also follow pathways of poor food choices and reduced physical activity-both of which can also have deleterious effects on health (10,25). Among the health consequences of following these pathways have been rapid increases in obesity and adult-onset diabetes (1 3,23,26). Members of this age group are influenced strongly by their peers, the media, and family situation and less by their knowledge of risky behaviors (6,21,22). Skipping meals is a common practice among adolescents: about 20 percent do not eat breakfast, and about half as many do not eat lunch (3,5,10,18,19). Skipping meals may lead to more snacking; for those who do not view skipping meals as a method of weight loss, snacks often compensate for missed calories and other key nutrients. The literature indicates that, on average, most children and adolescents average four eating occasions a day, with an upper range of 13 occasions among Mexican children who consumed as much as 45 percent of their energy from snacks (4,7,8). Research on the meal patterns of U.S. adolescents showed that most consume at least two meals (plus or minus snacks) on a consistent basis while some follow a highly inconsistent meal pattern: one meal and/or snacks all day (18). 15 Compared with adolescents with inconsistent meal patterns, those with consistent meal patterns consumed a diet that was adequate in calories and more nutrient dense (with respect to calcium, iron, vitamin E, and fiber) (1 1). Our study examines in more detail the types of food consumed by adolescents at each eating occasion and the nutrient contributions provided by each eating occasion to adolescents' total daily intakes. This study is unique: we examine snacking behaviors by using a nationally representative sample, and we determine the nutrient contributions of snacks. Previous studies have examined only the nutrient density of meals versus snacks without considering their contribution to the total diet, or previous studies have used very small samples to examine this research question (2,16). Methods Survey Design Food consumption data were provided by the 1989-91 Continuing Survey of Food Intakes by Individuals (CSFII), a survey conducted by the U.S. Department of Agriculture's Agriculture Research Service. A nationally representative sample was collected by using a multistage, stratified sample design of the 48 coterminous States and Washington, DC. Data were collected in four waves during each year: one in each season, between April 1989 and May 1991. In each wave, a different sample of participants was selected. The total number of participants in all age groups sampled was 15,192. Dietary data were collected for each individual in selected households. Using a 24-hour recall and two 1-day food records, individuals reported 3 consecutive days of intake. The female head of the household reported dietary intake for individuals less than 12 years 16 old. We were interested in the eating patterns of adolescents, thus our analysis was restricted to 11- to 18- year-olds who reported 3 days of dietary intake (n= 1,31 0). The classification of individuals into meal-pattern categories did not differ between 11- and 12-year-olds, and the differences in nutrient composition of reported intakes of 11- and 12-year-olds were similar in magnitude to the differences between 12- and 13-year-olds. Therefore, 11-year-olds were included in the analysis despite differences in methods of data collection for dietary intake. Variables Meal Paffems Survey data include descriptors of eating occasion (breakfast, lunch, dinner, supper, snack, brunch, and extended consumption) as well as the time of day each food was consumed. To identify meal patterns, we first developed clear and invariable terminology for eating occasions: Breakfast, lunch, brunch, dinner, or snack. Respondents provided the name for each meal. When respondents reported consuming either supper or dinner, the eating occasion was designated as dinner; when the respondent reported consuming both supper and dinner, dinner was designated as lunch and supper designated as dinner. This categorization was based on analysis of the data, which indicated that dinner was consumed primarily as an evening meal (85 to 87 percent between 4 and 8 p.m.). When both supper and dinner were consumed, dinner was the midday meal (56 to 69 percent betwee!l 11 a.m. and 3 p.m.) and supper was the evening meal (70 to 81 percent between 4 and 8 p.m.). Eating occasions for 1.3 percent of foods were unknown or identified as extended consumption and therefore not included in our analysis. Three meal-pattern categories were created based on their ability to provide meaningful comparison of eating behaviors: Consistent, moderately consistent, and inconsistent. These categories are mutually exclusive and include all possible combinations of eating occasions. Respondents with a consistent meal pattern (n=538) consumed two or three meals (plus or minus snacks) on all 3 days of reported intake. Those with a moderately consistent meal pattern (n=726) consumed two or three meals (plus or minus snacks) on 2 of the 3 days of reported intake. And respondents with an inconsistent meal pattern (n=46) consumed only one meal (plus or minus snacks) or snacks only on all 3 days of reported intake. Personal, Household, and Demographic Charaderistics Population characteristics available directly from the CSFII were age, gender, race, region of residence, supplement use, school attendance, educational and employment status of the female head of the household, income status, and household size. Our derived variables were consumption of school-based meals and single- versus dual-parent households. Respondents who reported consuming at least one school-based lunch per week were classified as consumers of school lunch. This method was repeated for school breakfast. Classification as a single- or dual-parent household was based on the presence of a male or female head of household or female and male heads of household, respectively. Nutrient and Food-Grouping System The nutrient database was provided by USDA Survey Nutrient Data Base, Release #7 and was developed for the 1991 CSFII. For this analysis, we used nutrient information provided as the total average intake or as the average percentage of the Recommended Dietary Allowances (RDA) for all nutrients (1 2) consumed over 3 days. Family Economics and Nutrition Review The age- and gender-appropriate RDA values were used to calculate the average percentage of the RDA consumed. Grams of food consumed at each eating occasion were calculated by using the University of North Carolina at Chapel Hill food-grouping system. This system disaggregates major USDA food groups into 56 more distinct nutrient-based groups based on the composition of fat and dietary fiber. The University of North Carolina at Chapel Hill 's food-grouping system covered all foods that respondents reported eating (14,1 5). I Statistical Methods We used Student's t test and a chisquare test to compare the sociodemographic characteristics among the groups based on their meal patterns. Statistical testing, however, was not performed on the proportion of the nutrients or the grams of food contributed by each meal. To do so would have required many comparisons, resulting in our having to use a very stringent p value. Hence our analysis is descriptive. The results provide estimates representative of the U.S. population in the coterminous 48 States. We weighted the statistics for nonresponse and corrected the standard errors for the complex multistage design. We used the STATA survey option that allows for the effects of the complex sample design (20). Results Sociodemogra phic Characteristics Forty-one percent of the adolescents had consistent meal patterns; only 4 percent had inconsistent meal patterns (table 1). The 15- to 18-year-olds were more likely to have inconsistent meal 1This information is available upon request. 2001 Vol. 13 No. I Table 1 . Descriptive characteristics of 15- to 18-year-olds and their households, by meal-pattern category, 1989-91 CSFII Meal pattern Sociodemographic Moderately characteristic Consistent consistent Inconsistent Sample 538 726 46 Percent Female 47.0 52.61 60.91,2 Black 14.1 19.01 23.91,2 Attends school 93.1 87.21 75.61 Single-parent household 27.3 31.41 34.81 Female head of household attended college 34.9 31.1 1 27.31,2 Female head of household has < 12 years of education 26.9 31.11 31.81 Region Northeast 18.0 17.2 10.91.2 Midwest 30.9 26.51 28.3 South 31.2 35.41 34.8 West 19.9 20.9 26.1 1,2 Mean (± S.D.) Percentage of poverty 329 (264) 326 (249) 330 (209) Household size 4.7 (1 .9) 4.2 (1.4)3 4.2 (1 .5)3'4 1 Significantly different from the consistent meal pattern, chi -square analysis, p < 0 .05. 2Significontly different from moderately consistent meal pattern, chi-square analysis, p < 0.05. 3Significantly different from consistent meal pattern, weighted t test, p < 0.05. 4Significantly different from moderately consistent meal pattern, weighted ttest, p < 0.05. patterns. The consistent meal-pattern category had a higher percentage of respondents who were male, white, and who attended school. Adolescents with a consistent meal pattern were less likely to be from a single-parent household and more likely to be from a household in which the female head attended college. Neither the mean percentage of poverty nor years of education of the female bead of household differed significantly by meal-pattern category. Nutrient Profiles Based on Adolescents' Meal Pattern Adolescents with a consistent or moderately consistent meal pattern consumed 37 to 38 percent of their total energy from dinner. For adoles-cents in the inconsistent group, 43 percent of their total energy was consumed at dinner. Even more important is the difference in the role of snacks in their diet. Snacks comprised about 23 percent of the total day's energy for those following an inconsistent meal pattern but only 11 to 16 percent for those following the other two meal patterns. In total, the dinner meal and total snacks together provided more than two-thirds of the day's total energy for adolescents with an inconsistent meal pattern. Differences by age group were noted (figs. 1 and 3). For instance, 11- to 14-year-olds in the inconsistent group obtained 24 percent of energy from lunch while older adolescents in the same group obtained 9 percent of 17 Figure 1. Distribution of energy obtained, by meal-pattern groups of 11-to 18-year-olds, 1989-91 CSFII Meal-pattern groups Consistent Cinner Total 38%-snacks 11% Brlllch 1% C3in8%ne r Moderately consistent 25% Total sna~s 16% Mlalng 1% Inconsistent Total lniOkl 1%'11 Figure 2. Distribution of energy obtained, by meal-pattern groups of 11-to 14-year-olds, 1989-91 CSFII Meal-pattern groups Consistent Moderately consistent Total Inconsistent Total .... ks 11% Lunoh 24'4 Figure 3. Distribution of energy obtained, by meal-pattern groups of 15-to 18-year-o/ds, 1989-91 CSFII Meal-pattern groups 18 Consistent Total ~~lllllllllfll~sn~a~s 12'1. 0.01% Breakfast 20'1. Moderately consistent Inconsistent Total Total 25% Family Economics and Nutrition Review energy from lunch. For II- to 14-yearolds with inconsistent meal patterns, breakfast provided 5 percent of energy, compared with almost twice that amount for the older adolescents. Brunch was more common for older adolescents with inconsistent meal patterns. Among older adolescents (15- to 18-year-olds) following an inconsistent meal pattern, dinner and snacks were far more important than they were for younger adolescents (II- to 14-year-olds). Most nutrients follow the same pattern as that for energy (table 2). Breakfast provided the same proportion of nutrients for adolescents with consistent and moderately consistent meal patterns; whereas, lunch appeared to have lower proportions of fat, protein, carbohydrates, calcium, fiber, and sodium for adolescents with moderately consistent meal patterns. The proportions of nutrients from brunch were very low (no more than 3.5 percent) for adolescents with the consistent and moderately consistent meal patterns but closer to I 0 percent for their counterparts with inconsistent meal patterns. Dinner provided similar proportions of nutrients for all three groups. For adolescents with inconsistent meal patterns, the proportion of nutrients provided by total snacks was nearly double the nutrients provided to adolescents with consistent meal patterns. The proportion of nutrients corning from snacks for the moderately consistent group falls between those of the other two groups. Food Consumed Based on Adolescents' Meal Patterns Interesting differences were noted in the types of foods consumed at each eating occasion across meal-pattern groups. At breakfast, adolescents with a consistent meal pattern, compared with adolescents in the other groups, had 200 I Vol. 13 No. I higher per capita consumption of both low- and medium-fat milk, egg items, low-fiber breads, cooked and ready-toeat cereals, high-fat desserts, and juices (fig. 4). In contrast, adolescents with moderate and inconsistent meal patterns consumed more soft drinks. At lunch, adolescents with consistent meal patterns consumed more milk and higher amounts of total poultry, high-fat desserts, vegetables, fruits, and high-fat grain-based mixed dishes (pizza and macaroni and cheese, etc.), compared with other adolescents (fig. 5). There was no difference in beef! pork consumption between adolescents with consistent and moderately consistent meal patterns, which were, however, higher than that for adolescents with inconsistent meal patterns. The inconsistent and consistent groups had the same quantity of high-fat potato consumption. For dinner, teens with an inconsistent meal pattern had higher intakes of poultry, green/orange vegetables, highfat grain-based mixed dishes, high-fat breads, and soft drinks and a lower intake of low- and medium-fat milk, soy and legumes, and fruits, compared with their other adolescent counterparts (fig. 6). In contrast, snacks for the inconsistent group contained more grams per capita of milk items, in particular medium-fat milk items (whole milk and milk shakes) and soft drinks than was the case for the consistent group (fig. 7). All three groups of adolescents had similar intakes of fruits, high-fat desserts, high-fat salty snack items (chips, salty crackers, etc.) and high-fat grain-based mixed dishes. Patterns within each group revealed that the amount of soft drink consumed per capita at lunch, dinner, and total snacks was higher than the amount of milk consumed (figs. 4-7). High-fat, low-fiber bread is more commonly eaten at breakfast, compared with other bread options available. Low-fiber In total, the dinner meal and total snacks together provided more than twothirds of the day's total energy for adolescents with an inconsistent meal pattern. 19 Table 2. Proportion of nutrients provided by each meal among 11- to 18-year-olds, by meal-pattem consumption 1 Consistent meal pattern Moderately consistent meal pattern Inconsistent meal pattern Nutrient Mean Standard deviation Mean Standard deviation Mean Standard deviation Breakfast Energy 19.23 9.69 18.24 10.62 7.94 8.51 Fat 16.07 11.76 15.66 11 .56 6.31 7.94 Saturated fat 17.92 12.87 17.03 12.20 7.67 8.79 Protein 17.27 9.95 16.15 10.60 9.63 11.86 Carbohydrate 21.95 10.28 20.81 11.93 8.70 9.38 Calcium 28.94 16.53 26.31 16.81 15.81 18.04 Cholesterol 23.35 20.16 21 .35 20.82 10.46 17.45 Iron 29.24 17.26 27.47 19.01 19.72 25.64 Folate 39.11 19.16 35.90 22.16 22.66 25.07 Zinc 20.70 13.16 18.63 15.20 10.51 14.33 Fiber 14.68 10.40 16.51 13.05 7.92 12.93 Sodium 16.23 9.54 16.17 10.40 8.86 10.52 Brunch Energy 0.47 3.48 2.11 5.80 8.58 13.69 Fat 0.51 3.92 2.44 7.85 9.82 16.16 Saturated fat 0.58 3.93 2.47 7.88 9.07 15.41 Protein 0.45 3.16 2.05 6.84 7.02 11.39 Carbohydrate 0.46 3.60 2.04 5.39 7.97 12.99 Calcium 0.60 3.76 2.26 7.36 8.05 16.38 Cholesterol 0.39 2.88 3.50 12.84 7.97 12.93 Iron 0.36 2.52 2.12 6.87 8.00 12.84 Folate 0.40 2.69 2.41 7.76 9.30 16.49 Zinc 0.43 3.04 1.87 5.76 8.65 15.52 Fiber 0.57 5.12 2.13 6.84 8.14 15.44 Sodium 0.39 2.88 2.11 6.30 9.89 17.09 Lunch Energy 31.22 11 .64 24 .52 12.58 11.94 18.98 Fat 33.29 14.11 26.10 14.56 12.52 20.91 Saturated fat 32.43 14.53 26.33 15.62 11 .94 19.99 Protein 30.46 12.04 24.69 14.14 10.32 18.14 Carbohydrate 30.03 12.02 23.48 12.39 12.39 19.05 Calcium 30.46 15.74 24 .82 16.11 10.32 19.26 Cholesterol 26.65 14.67 22.49 17.68 10.41 18.49 Iron 26.36 12.03 21 .56 13.00 9.54 16.59 Folate 22.88 12.70 19.90 14.38 8.88 18.33 Zinc 27.74 12.99 23.75 14.88 10.42 18.25 Fiber 33.16 15.87 25.26 14.25 12.65 22.14 Sodium 32.26 12.70 25.06 14.12 11.52 20.20 1 The percentage may not total to 1 00 for each nutrient because of rounding and the sm~ll percentage of foods with a missing eating-occasion classification. (Continued) 20 Family Economics and Nutrition Review Table 2. Proportion of nutrienfs provided by each meal among 11- to 18-year-olds, by meal-pattem consumption 1 (continued) Consistent meal ~attern Moderately consistent meal ~attern Inconsistent meal pattern Nutrient Mean Standard deviation Mean Standard deviation Mean Standard deviation Dinner Energy 37.72 10.53 38.01 14.60 43.21 26.07 Fat 40.17 13.68 39.79 16.77 46.58 26.58 Saturated fat 38.54 14.52 37.87 17.10 44.04 28.06 Protein 45.23 12.33 45.90 16.52 55.07 25.52 Carbohydrate 33.61 10.92 33.92 14.65 37.49 26.89 Calcium 29.86 14.20 30.84 17.28 42.09 29.05 Cholesterol 43.29 18.86 40.97 20.76 49.50 28.39 Iran 36.81 12.84 37.00 17.46 44 .53 30.34 Folate 29.93 13.59 30.53 16.87 38.64 26.11 Zinc 43.80 14.72 43.86 18.95 49.59 28.25 Fiber 41.84 15.89 41 .61 18.22 46.14 27.62 Sodium 44.83 12.15 44.52 16.40 52.51 26.52 Total snacks Energy 10.92 9.69 15.50 13.74 23.16 19.02 Fat 9.65 10.06 14.40 15.01 21.41 20.34 Saturated fat 10.16 10.94 14.78 15.44 23.38 21 .52 Protein 6.38 7.17 9.85 11 .30 15.36 15.93 Carbohydrate 13.35 11 .53 18.06 15.16 26.91 20.41 Calcium 9.71 11 .97 14.24 15.90 19.52 17.71 Cholesterol 6.05 8.59 10.09 13.08 17.83 21.74 Iron 7.06 8.62 10.45 11.87 15.82 17.82 Folate 7.48 9.93 9.79 12.26 18.75 19.36 Zinc 7.09 8.33 10.53 11 .81 17.37 17.92 Fiber 9.39 10.96 13.11 13.41 22.09 19.26 Sodium 6.12 6.92 10.74 13.18 14.26 14.47 1 The percentage may not total to 100 for each nutrient because of rounding and the small percentage of foods with a missing eating-occasion classification. cereal is consumed more than high fiber, and citrus juices are consumed more than noncitrus juices or fruit drinks at breakfast. At lunch, medium fat beef/pork and poultry were eaten more than the low- or high-fat option. The grams per capita for luncheon meats was equally distributed among each fat option. In contrast to breakfast, the low-fat, low-fiber bread option and fruit drinks were eaten in greater amounts at lunch for the consistent and moderately consistent 2001 Vol. 13 No. I groups. And for all three groups, there was a higher per capita consumption of the high-fat versus the low-fat of grain-based mixed dishes. The type of bread consumed at the dinner meal was similar to that seen at breakfast. And the higher fat version of grain-based meals was once again consumed more than the low-fat version at dinner for all three groups. Patterns within groups for the different types of foods consumed as snacks were similar. Discussion Dietary intake patterns of U.S. adolescents are poor. Skipping meals, excessive snacking, and consumption of excessive high-fat, poor nutritionally dense foods are many of the issues raised in the literature. However, few studies have used nationally representative samples to examine the meal and food patterns of U.S. adolescents. This study highlights the large variation in eating patterns among U.S. adolescents. 21 Figure 4. Grams consumed at breakfast by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups1 Figure 5. Grams consumed at lunch by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups1 Breakfast food items (grams) Low-lot milk Medium-lot milk Eggs and egg dishes Low-lot/low-fiber bread Low-lot/high-fiber bread High-fat/low-fiber bread High-fat desserts Posta, rice, and cooked cereals Ready-to-eat cereals Citrus fruits and juices Other fruits and juices Sugars and jellies Coffee and teo Soft drinks Fruit drinks 0 20 40 60 • Consistent 70 68 80 85 8 100 120 Lunch food items (grams) Low-fat beef and pork Medium-fat beef and pork Low-fat poultry Medium-fat poultry Low-/medium-fat luncheon meats/hot dogs High-fat luncheon meats/ hot dogs High-fat desserts High-fat grain-based mixed dishes Citrus fruits and juices Other fruits/juices High-fat potatoes Vegetables Soft drinks 88 4 0 20 40 60 80 100 0 Moderately consistent • Inconsistent • Consistent D Moderately consistent • Inconsistent 1 University of North Carolina at Chapel Hill food-grouping system. 1 University of North Carolina at Chapel Hill food-grouping system. In particular, we show that teens differ markedly by the proportion of food intake from each meal and the types of foods eaten, based on consistent, moderately consistent, or inconsistent meal patterns. Regarding snacks, our results differ from the few published studies on this topic. Our study fmds that snacks contribute much less to the total diet than reported previously. For most adolescents (97 percent), meals contribute, on average, 20 to 40 percent 22 of the total day's energy, compared with 10 to 15 percent contributed by snacks. One study has found that about 25 to 33 percent of the total day's energy comes from snacks (16). Other publi~hed studies have focused more on the frequency of snacking and the snack foods adolescents like to eat (4,5,8). Our study shows that for all the adolescents, a higher proportion of the total day's intake of fat is consumed at dinner. Otherwise, meals and snacks provide similar proportions of the other macronutrients. Our results regarding macronutrients offer a different view of examining macronutrient intake; others who have examined the nutrient density of meals and snacks have found that meals are higher in fat and lower in carbohydrates than are snacks (8). The nutrient contribution of snacks is more significant for those adolescents following an inconsistent meal pattern, compared with adolescents following Family Economics and Nutrition Review Figure 6. Grams consumed at dinner by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups' Figure 7. Grams consumed at snack by adolescents following a consistent, moderately consistent, or inconsistent meal pattern, by selected food groups' Dinner food items (grams) Soy and legumes High-fat/ low-fiber bread Green and orange vegetables Citrus fruits/ juices Other fruits Diet soft drinks 0 20 5 40 60 80 100 120 • Consistent Snack food items {grams) Low-fat milk Medium-lot milk Low-fat desserts High-fat desserts High-fat, salty snacks High-fat groin-based mix Citrus fruits/ juices Other fruits Candy Fruit drinks Diet soft drinks 0 20 40 60 80 100 120 140 160 180 • Consistent 0 Moderately consistent • Inconsistent D Moderately consistent • Inconsistent 1 University of North Carolina at Chapel Hill food-grouping system. 1 University of North Carolina at Chapel Hill food-grouping system . other meal patterns. This occurs simply by how this group was defmed, those consuming one meal plus or minus snacks on 3 days of intake. The nutrient contribution of snacks for this adolescent group is similar to that reported by Ruxton et al. (1 6) in a study of 7- to 8- year-olds (n=136), from five schools in Scotland). In our study, snacks for the inconsistent group provided more of most nutrients than did breakfast or lunch with the exception of iron and folate, which were higher at breakfast. By using the 1977 Nationwide Food Consumption Survey (NFCS), 2001 Vol. 13 No. I researchers found that for most adolescents, snacks compared with meals contributed significantly more magnesium, calcium, vitamin A, and vitamin C to the diet (2). For the only nutrient on which we overlapped, calcium, this was not found in the 1989-91 USDA survey. Because of the frequency of snacking and the significant proportion of energy and other nutrients that snacks provide adolescents with an inconsistent meal pattern, we believe the nutritional quality of snacks has important implications for the health status of these adolescents. Another nutritionally important issue is adolescents' high intake of soft drinks and lower intake of milk [also noted in the 1977 NFCS survey (9)]. These consumption patterns apply to adolescents regardless of meal-pattern group. Even though adolescents with a consistent meal pattern consume the most milk, their calcium intakes are lower than recommended. Also, adolescents appear to be consuming more high-fat and low-fiber foods than the more healthful alternatives. Consuming more high-fat and low-fiber foods may have serious health consequences (i.e., 23 obesity, osteoporosis, and cardiovascular diseases) if they are consumed in high amounts throughout life. The reasons for the high consumption of these types of foods may be directly related to their source (home vs. awayfrom- home food sources) as well as the taste preferences of adolescents. A Minnesota survey of 900 adolescents reported a strong preference for highfat foods that related to taste appeal despite the health consequences associated with consumption of these foods (21). A limitation of this study is the small sample size for the group with an inconsistent meal pattern. However, there were 27 million adolescents in the United States (24) around the time of this survey. Thus 950,000 U.S. adolescents are represented as following an inconsistent meal pattern for 3 days. In general, adolescents are consuming a large quantity of carbonated beverages and few fruits and vegetables. And for adolescents who follow an inconsistent meal pattern, dinner and snacks provide a disproportionate amount of nutrients. Differences are also noted in food selection: adolescents following an inconsistent meal-pattern group consume more types of fast foods . Both meal-pattern and food-selection behaviors should be used to target future public health messages to adolescents. More research is warranted on the determinants of adolescent eating patterns. Information on the determinants could help guide interventions for changing eatingpattern behaviors noted in this study. 24 Acknowledgments We thank the Nestle Research Center for providing financial support for this study. We thank Dr. Henri Dirren for initiating the collaboration between Nestle and the University of North Carolina at Chapel Hill. We also thank Dan Blanchette, Terri Carson, Claire Zizza, Lynn Igoe, and Frances Dancy for their assistance. Family Economics and Nutrition Review References I. Berenson, G.S., McMahan, C.A., Voors, A.W., Webber, L.S., Srinivasan, S.R., Frank, G.C., Foster, T.A., and Blonde, C.V. 1980. Cardiovascular Risk Factors in Children, the Early Natural History of Atherosclerosis and Essential Hypertension. Oxford University Press, New York. 2. Bigler-Doughten, S. and Jenkins, R.M. 1987. Adolescent snacks: Nutrient density and nutritional contribution to total intake. Journal of the American Dietetic Association 87:1678-1679. 3. Cavadini, C. 1996. Dietary habits in adolescence: Contributions of snacking. In A. Ballabriga (Ed.), Feeding from Toddlers to Adolescence. Nestle Nutrition Workshop Series vol. 37. Lippincott-Raven, Philadelphia. Nestle Nutrition Services, Vevey, Switzerland. 4. Cross, A.T., Babicz, D., and Cushman, L.F. 1994. Snacking patterns among 1800 adults and children. Journal of the American Dietetic Association 94: 1398-1403. 5. Devaney, B.L., Gordon, A.R., and Burghardt, J.A. 1995. Dietary intakes of students. American Journal of Clinical Nutrition 6l(supp 1 ):205S-212S. 6. Friedman, H.L. 1989. The health of adolescents: Beliefs and behavior. Social Science and Medicine 29:309-315. 7. Garcia, S.E., Kaiser, L.L., and Dewey, K.G. 1990. The relationship of eating frequency and caloric density to energy intake among rural Mexican preschool children. European Journal of Clinical Nutrition 44:381-387. 8. Gatenby, S.J. 1997. Eating frequency: Methodological and dietary aspects. British Journal of Nutrition 77(supp l):S7-S20. 9. Guenther, P.M. 1986. Beverages in the diets of American teenagers. Journal of the American Dietetic Association 86:493-499. 10. Lund, E.K., Lee-Finglas, W.E., Southon, S., Gee, J.M., Johnson, I.T., Finglas, P.M., and Wright, A.J. 1992. Dietary fat intake and plasma lipid levels in adolescents. European Journal of Clinical Nutrition 46:857-864. II. Miller, E. C. and Maropis, C. G. 1998. Nutrition and diet-related problems. Primary Care 25:193-210. 12. National Research Council, Subcommittee on the Tenth Edition of the RDAs, Food and Nutrition Board. 1989. Recommended Dietary Allowances (I Oth ed.) National Academy Press, Washington, DC. 13. Pinhas-Hamiel, 0. and Zeitler, P. 1996. Insulin resistance, obesity, and related disorders among black adolescents. Journal of Pediatrics 129(3):319-320. 14. Popkin, B.M., Haines, P.S., and Reidy, K.C. 1989. Food consumption trends of U.S. women: Patterns and determinants between 1977 and 1985. American Journal of Clinical Nutrition 49:1307-1319. 15. Popkin, B.M., Siega-Riz, A.M., and Haines, P.S. 1996. A comparison of dietary trends between racial and socioeconomic groups in the United States. New England Journal of Medicine 335:716-720. 2001 Vol. 13 No.1 25 16. Ruxton, C.H.S., Kirk, T.R., and Belton, N.R. 1996. The contribution of specific dietary patterns to energy and nutrient intakes in 7-8 year old Scottish schoolchildren. III. Snacking habits. Journal of Human Nutrition and Dietetics 9:23-31. 17. Serdula, M.K., Ivery, D., Coates, R.J., Freedman, D.S., Williamson, D.F., and Byers, T. 1993. Do obese children become obese adults? A review of the literature. Preventive Medicine 22:167-177. 18. Siega-Riz, A.M., Popkin, B.M., and Carson, T. 1998. Three squares or mostly snacksWhat do teens really eat?: A sociodemographic study of meal patterns. Journal of Adolescent Health 22:29-36. 19. Siega-Riz, A.M., Popkin, B.M., and Carson, T. 1998. Trends in breakfast consumption for children in the U.S. from 1965-1991. American Journal of Clinical Nutrition 67:748s- 756s. 20. STATA Corporation. 1997. Stata statistical software: Release 5.0. STATA Corporation, College Station, TX. 21. Story, M. and Resnick, M.D. 1986. Adolescents' view on food and nutrition. Journal of Nutrition Education 18: 188-192. 22. Sweeting, H., Anderson, A., and West, P. 1994. Socio-demographic correlates of dietary habits in mid to late adolescence. European Journal of Clinical Nutrition 48:736-748. 23. Troiano, R.P., Flegal, K.M., Kuzmarski, R.J., eta!. 1995. Overweight prevalence and trends for children and adolescents. The National Health and Nutrition Examination Surveys. Archives of Pediatrics and Adolescent Medicine 149: I 085-1091. 24. U.S. Department of Commerce, Economics and Statistics Administration, Bureau of the Census. 1992. 1990 Census of the Population: General Population Characteristics. 25. U.S. Department of Health and Human Services, Public Health Service, Office of Disease Prevention and Health Promotion, Centers for Disease Control, and National Institute on Drug Abuse. 1989. The National Adolescent Student Health Survey: A Report on the Health of Americas Youth. Third Party Publishing Co., Oakland, CA. 26. U.S. Department of Health and Human Services. 1996. Update: Prevalence of overweight among children, adolescents, and adults-United States 1988-94. Morbidity and Mortality Weekly Report 46:199-202. 26 Family Economics and Nutrition Review Rachel K. Johnson, PhD, MPH, RD The University of Vermont Celeste V. Panely, MS, RD The University of Vermont Min Qi Wang, PhD The University of Maryland 2001 Vol. 13 No. 1 Associations Between the Milk Mothers Drink and the Milk Consumed by Their School-Aged Children The declining milk intakes of U.S. children are of concern because milk is the primary source of calcium in children's diets. The aim of th is study was to determine the predictors of milk consumption in U.S. schoolaged children (ages 5- 17) by using dietary intake data from the USDA 1994-95 Continuing Survey of Food Intakes by Individuals (CSFII). Sociodemographic variables, type of milk consumed (skim, 1 %, 2%, whole, or none), and mothers' milk intake (type and amount) were examined as possible predictors. The sample consisted of 1,303 CSFII participants. Sample weights were applied to allow for generalizations to the entire U.S. school-aged population. Children's average milk intake was 300.4 grams per day. For every gram of milk a mother consumed, her child's intake increased by 0 .64 grams. Two percent milk was the most commonly consumed milk among the children. For each type of milk consumed by mothers, children were at least 30 times more likely to drink that same type. The strong association between the milk consumed by mothers and the amount and type of milk consumed by U.S. school-aged children should be considered when designing intervention programs aimed at increasing children 's milk intake. E vidence suggests that attainment of peak bone mass by early adulthood may be the most effective protection against osteoporotic fractures later in life (23). Throughout the developmental years, adequate calcium intake is essential to support bone growth (16). Substantial evidence exists linking higher calcium intakes with improved skeletal health in children (2,3,16,21,23,30). Data from the U.S. Department of Agriculture's (USDA) nationwide food consumption surveys reveal that most U.S. schoolaged children have calcium intakes that are below recommended levels (4). Calcium intake is especially problematic for girls, with 59 percent ages 6-11 and 86 percent ages 12-18 not meeting recommendations (4) . Milk and dairy products are the primary source of calcium in children 's diets (8). Johnson and colleagues found that in a large sample of school-aged children, on average, only those children who consumed milk at the noon meal met their daily requirement for calcium (1 5). Rising consumption of soft drinks has been shown to have a negative effect on calcium intake among children and adolescents by 27 competing with milk as a preferred beverage (9). On the other hand, whole and 2% milk are leading sources of fat and saturated fat in the diets of U.S. children (33). USDA food consumption survey data indicate that for children in all age groups, mean total and saturated fat intakes exceed the recommended levels (4). Because milk is an important contributor of both calcium and fat in the diets of children, it is important to identify the predictors of children's milk intake (both type and amount). The aim of this study was to identify predictors of U.S. school-aged children's milk intake. Familial aggregation studies show similarities in nutrient intake between parents (especially mothers) and their children (26). Hence, milk consumption patterns of mothers were included, along with sociodemographic variables, in the research model as possible predictors of children's milk intake. Findings from this study will assist nutrition policymakers, school nutrition personnel, school administrators, nutrition educators, and parents in developing appropriate intervention strategies to address the problem of children's declining milk consumption. Methods Sample The research sample was obtained from the 1994-95 USDA Continuing Survey of Food Intakes by Individuals (CSFII). The CSFII is a continuing component of the USDA Nationwide Food Consumption Survey. The surveys provide data on demographics as well as dietary intake for a nationally representative sample of noninstitutionalized persons residing in the United States. The 1994-95 survey included data on the food and nutrient intakes of 5,598 individuals. The response rate of the 28 survey was 80 percent for Day I dietary intake data and 76 percent for Day 2 (4). These response rates are acceptable by research standards (7) . Trained interviewers used the multiplepass 24-hour recall method to collect 2 days of dietary intake data from each respondent. The multiple-pass 24-hour recall method has been validated as an accurate measure of children's dietary intake (11). All children ages 5 to 17 years with 2 complete days of dietary intake data (N= 1 ,303) and their mothers were included in this study. Study Variables The study investigated predictors of both the amount and type (skim, 1%, 2%, whole, or none) of milk consumed by U.S. school-aged children. The following sociodemographic variables were assessed as possible predictors: Child gender, age, and race; household income; geographic region; urbanization; and mother's age, education, and occupation. Participation in the USDA Food Stamp Program and participation in the USDA national school lunch and school breakfast programs were also included as possible predictors of a child's consumption of milk. Milk is required to be served in the national school lunch and school breakfast programs (5) . Mothers' milk consumption patterns (both type and amount) were included as potential predictors. A mother's nutrient intake has been shown to influence her child's nutrient intake (26) . In addition, studies by Pelletier and colleagues indicated that.among adult milk drinkers, consumption of lower fat versions of milk (I% and skim) was associated with increased average daily milk consumption (2 7). If the same is true for children, promotion of I% and skim milk in this population could have a positive influence on calcium intake. The dependent variables in the analysis were "Child Milk Amount" and "Child Milk Type." Child Milk Amount was defined as the 2-day mean intake in grams of fluid milk consumed by the sample child. The 7,250 food codes in the CSFII database were searched, and all codes whose primary ingredient was fluid cows' milk were included. Items such as flavored milk, evaporated milk, dry reconstituted milk, eggnog, and milk shakes were included. However, items such as flavored drinks (e.g., Yoo-hoo®l, canned meal replacements (e.g., Instant Breakfast®l, and infant formulas were excluded. Child Milk Type was defined as the type of milk (skim, 1%, 2%, whole, or none) most often consumed by the sample child. The CSFII food codes were searched and all fluid milks were grouped into one of the four categories: Skim, I%, 2%, or whole. For example: "milk, chocolate, skim milk based" was categorized as skim; "milk, dry, reconstituted, whole" was categorized as whole. The category consumed in the greatest quantity in grams over 2 days by each sample child was considered the Child Milk Type. Statistical Analysis The Statistical Export and Tabulation System (SETS) software and the Statistical Analysis System (SAS) were used to format and recode the data for statistical analysis. Statistical significance was set at p S 0.05 for all analyses. To compensate for variable probabilities of selection, differential nonresponse rates, and sampling frame considerations, we applied sample weights in both the descriptive and comparative analyses. The Survey Data Analysis System (SUDAAN) was used to weight the sample, compute variances, and run the statistical procedures. Applying sample weights aiiows the findings to be generalized to the entire U.S. population of school-aged children. Analysis of variance and analysis of Family Economics and Nutrition Review covariance were used to determine both the bivariate and multivariate effect of each independent variable on the dependent variable, Child Milk Amount. Only those independent variables that were significant at the bivariate level were included in the final multivariate model. Chi-square statistics were used to identify independent variables associated bivariately with Child Milk Type. The Multinominal Logistic Model was used for the multivariate analysis of Child Milk Type. As with the Child Milk Amount model, only those independent variables that were significant at the bivariate level were included in the multivariate model. 1 The results of the multinomial model were presented as odds ratios, which describe the change in likelihood of one outcome (e.g., drinking whole milk) versus another outcome (e.g., drinking 2% milk) given a particular characteristic or level of predictor (e.g., being a male compared with being a female) (31). In multinominallogistic models, each outcome (skim, 1%, whole, none) is compared with a reference category, which we determined to be 2% milk-the most common type of milk consumed. Odds ratios greater than 1.0 indicate an increased likelihood of consumption of that type of milk (compared with 2%) for children with that characteristic; whereas an odds ratio of less than 1.0 indicates a lower likelihood of consuming that type of milk (compared with 2%) for children with that characteristic. Both unadjusted and adjusted odds ratios were calculated. 1The Multinominal Logistic Model is an extension of the logistic regression model. While logistic models can only process dichotomous outcome variables, the multinominal model can include outcomes with two or more categories (25). 2001 Vol. 13 No. 1 Table 1. Amount and type of milk consumed1 by children ages 5-17 who provided 2 days of dietary intake data, 1994-95 CSFII Type of milk consumed Percent Mean amount (grams) Skim 11.4 376.62 1% 9.6 407.9 2% 32.0 385.4 Whole 28.4 347.8 None 18.6 0.0 1Two-day mean intake of milk (groms/ day) = 300.4+ 11 .9 . 2There was no association between type (skim, 1%, 2%, whole) and amount of milk consumed. N= 1,303. This allows for the examination of the influence the independent variables have on the dependent variable (Child Milk Type) both before and after the model is adjusted for all the covariates. Any odds ratio with 95 percent confidence intervals that included 1.0 was not considered statistically significant. Results Demographics The unweighted sample of CSFII respondents consisted of 1,303 participants. The children's average age was 11.6 years; the mothers', 39 years. Most of the sample was white, and was divided relatively equally between boys and girls. The sample was geographically diverse and representative of the U.S. population. Most participants resided in suburban areas, and the average yearly household income was about $44,000. The mothers' most common classes of occupation included professionaVtechnical and clericaV sales. Twenty-four percent of the children were eligible to receive free or reduced-price lunches, and 14 percent were eligible to receive free or reduced-price breakfasts. Milk Consumption The 2-day mean milk intake for children was 300.4 grams per day (table 1). Mothers' mean intake was 109.0 grams per day. Of the types of milk consumed by children (skim, l %, 2%, whole, and none), 2% milk was most commonly consumed, followed by whole milk. Two percent milk was also the most commonly consumed type by mothers, followed closely by whole milk. No significant associations were found between the type (skim, I%, 2%, or whole) and amount of milk consumed by children. Predictors of the Amount of Milk Consumed by Children Based on the bivariate analysis, the type and amount of milk consumed by mothers, geographic region, and the child's gender were associated with Child Milk Amount. Hence, these variables were entered into the multivariate model. In this model, the type of milk mothers consumed was not significant; however, geographic region, the child's gender, and the amount of milk mothers consumed each had a significant effect on the amount of milk consumed by children. In the multivariate analysis, children from the Midwest had significantly higher milk intakes than children from 29 For every 1 gram of milk a mother consumed, her child's intake increased by 0.64 grar:ns. 30 Table 2. Amount of milk consumed by children ages 15-17: Analysis of covariance (ANCOVA)1 of significant relationships, 1994-95 CSFII Variable Mothers' milk intake (milk type) Skim 1% 2% Whole None Mothers' milk intake (milk amount, grams) Region Northeast Midwest West South Child's gender Mole Female Beta coefficient (±SE Beta) 1.94±35.1 28.3 ± 28.6 4.3 ± 32.0 29.0 ± 31.7 0.00 ± 0.00 0.64 ± 0.1 11.6 ± 26.9 71.8 ± 35.6 49.3 ± 29.7 0.0 ± 0.0 120.0 ± 16.9 0.0 ± 0.0 P-volue 0.96 0.33 0.89 0.37 <0.001 0.66 0.05 0.10 <0.001 1 F value for overall model = 137.07; P-value for model < .001; Intercept = 145.09. - = No reference category. N=1 ,303. the South (table 2). Boys in the sample consumed 120 grams more milk per day than girls consumed. Maternal milk intake was significantly and positively associated with the amount of milk children consumed. For every 1 gram of milk a mother consumed, her child's intake increased by 0.64 grams. Predictors of the Type of Milk Consumed by Children Of the 12 independent variables, the children's age, gender, and race; geographic region; eligibility for free and reduced-price school lunch and breakfast; mothers' age and level of education; and the amount and type of milk consumed by mothers had a significant bivariate effect on Child Milk Type. Urbanization and participation in the Food Stamp Program were not significant predictors, and were therefore dropped from the multivariate model. In the multivariate model, older children were more likely to drink skim milk or no milk than were younger children (table 3). Children who paid full price for lunch were more likely to drink skim milk, compared with children who were eligible to receive (and presumably received) free school lunch. Children from the Northeast were more likely than children from the South to drink 1% milk; whereas, children from the South were more likely than children from the Midwest to drink whole milk or no milk. Black children were more likely to drink whole milk or no milk than were White children. Girls were twice as likely to drink no milk, compared with boys. The type of milk mothers drank was a very strong predictor of the type of milk children drank. Two percent milk was used as the reference category for Child Milk Type, because this was the type most commonly consumed by the Family Economics and Nutrition Review Table 3. Milk consumed by children ages 5-17: Results of unadjusted and adjusted odds ratios/ 1994-95 CSFII CHILD Age (years) 13-17 9-12 5-8 Race Black White Other Gender Female Mole School lunch None Free Reduced Full School breakfast None Free Reduced Full MOTHER Age (years) 40-60 30-39 20-29 Education College graduate Some college High school Type of milk consumed Skim 1% Whole None 2% Amount of milk consumed (grams) > 360 241-360 121-240 0-120 REGION t. Northeast Midwest West South Unadj OR2 1.8 1.3 1.0 0.1 1.0 0.5 1.4 1.0 1.3 0.1 0.4 1.0 2.7 0.8 0.0 1.0 2.4 1.4 1.0 2.1 2.0 1.0 37.7 4.3 4.6 7.2 1.0 1.0 1.9 0.6 1.0 1.3 0.6 0.7 1.0 Skim Adj OR 2.2 1.6 1.0 0.4 l.O l.O 1.5 1.0 1.0 0.2 1.2 1.0 2.8 8.6 0.4 1.0 0.7 0.8 1.0 2.0 1.5 1.0 30.0 4.7 5.9 7.2 1.0 0.9 1.4 0.6 1.0 1.5 0.8 1.0 1.0 1.2, 3.9 0.9, 2.9 0.1' 1.2 0.3, 3.8 0.9, 2.7 0.4, 2.8 0.1 , 0.4 0.2, 7.4 0.5, 15.5 1.1' 69.8 0.1, 3.4 0. ' 3.9 0.1 , 4.5 0.5, 8.0 0.7, 3.1 9.4, 95.8 0.8, 28 .3 1.5, 23 .4 2.0, 26.1 0.2, 3.2 0.5, 4.3 0.1, 2.4 0 0.7, 3.2 0.3, 2.0 0.4, 2.2 Unadj OR 0.7 0.7 1.0 0.1 l.O 0.7 0.9 1.0 3.6 1.1 0.7 l.O 1.4 1.0 1.3 1.0 2.1 2.3 1.0 3.6 2.2 1.0 3.2 67.2 2.8 3.1 l.O 1.6 1.8 0.9 l.O 4.5 0.8 2.0 1.0 1% Adj OR 1.2 0.9 1.0 0.5 1.0 0.4 1.2 1.0 2.3 1.3 0.9 1.0 0.7 1.9 2.6 1.0 1.7 2.8 1.0 2.5 1.6 1.0 3.7 114 2.2 4.2 1.0 95% Cl 0.6, 2.6 0.4, 1.7 0.2, 1.6 0.1' 1.5 0.7, 2.2 1.2, 4.4 0.3, 4.9 0.2, 4.6 0.1 , 4.2 0.2, 21 0.3, 27 0.4, 7.1 0.7, 11 0.9, 6.8 0.8, 3.3 1.1' 12 31, 416 0.5, 8.5 1.2, 15 1.3 0.3, 6.0 2.8 0.6, 12 1.0 0.3, 2.8 1.0 5.5 1.3, 23 0.9 0.2, 3.3 4.0 1.0, 16 1.0- Unadj OR 0.8 0.8 1.0 6.4 1.0 2.7 0.9 1.0 1.2 2.6 2.0 1.0 3.0 7.3 7.9 1.0 0.4 0.7 1.0 0.1 0.4 1.0 3.7 5.2 50.1 10.8 1.0 1.0 1.3 0.5 1.0 l.l 0.3 0.7 1.0 Whole Adj OR 1.2 1.1 1.0 3.3 1.0 l.7 0.7 1.0 0.9 0.8 1.4 1.0 3.9 3.6 3.1 1.0 0.9 1.5 1.0 0 .3 0 .5 1.0 4 .1 8.2 45.8 13.0 1.0 2.0 1.4 0.9 1.0 1.1 0 .4 0.5 1.0 95% Cl 0.7, 2.1 0.7, 1.7 1.7, 6.4 0.6, 4.6 0.4, 1.1 0.5, 1.7 0.3, 2.6 0.4, 4.2 0.8, 18.5 0.6, 23 .9 0.4, 22.9 0.4, 2.2 0.6, 3.9 0.1' 0.7 0.3, 0.8 1.2, 14.6 2.6, 25.8 17.1 , 122.8 6.0, 28 .1 0.8, 5.1 0.4, 4.3 0.4, 2.3 0.4, 2.8 0.2, 0.7 0.2, 1.3 Unodj OR 3.8 1.0 1.0 3.2 1.0 0.8 2.1 1.0 1.4 0.6 0.8 1.0 1.9 1.8 2.7 1.0 2.7 1.4 1.0 1.2 0.8 1.0 4.5 8.1 7.4 6.9 1.0 0.4 0.2 0.3 1.0 0.8 0.3 0.5 1.0 None Adj OR 3.9 l.O l.O 3.0 1.0 1.3 2.1 1.0 1.0 0.3 0.8 1.0 1.8 3.4 3.0 1.0 1.2 1.0 1.0 1.3 0.8 1.0 4.3 8.6 5.9 3.8 1.0 0.7 0.3 0.5 1.0 0.9 0.4 0.7 1.0 95% Cl 2.2, 7.1 0.5, 1.9 1.2, 7.6 0.4, 4.1 1.1' 4.0 0.6, 1.9 0.1' 1.1 0.3, 2.4 0.4, 8.6 0.4, 26.7 0.4, 21 0.4, 3.8 0.3, 3.0 0.4, 3.8 0.4, 1.4 1.4, 13 3.9, 19 1.8, 19 1.6, 9.2 0.3, 2.0 0.1' 1.1 0.2, 1.3 0.4, 2.1 0.2, 0.8 0.3, 1.7 1Consumption of skim, 1%, whole, or no milk is compared to 2% milk, as 2% is the most commonly consumed milk type among both sample children and mothers. Odds ratios whose confidence limits do not include 1.0 are bolded. 20dds ratios. 3Confidence intervals. -=No reference category. N=l ,303. 2001 Vol. 13 No. 1 31 For each type of milk ... consumed by mothers, children were at least 30 times more likely to drink the some milk type as their mothers. 32 sample. For each type of milk (skim, I%, whole, or none) consumed by mothers, children were at least 30 times more likely to drink the same milk type as their mothers. In addition, the more educated a mother was, the less likely her child was to drink whole milk. The odds ratios for school breakfast, mother milk amount, and mothers' age were not significant; the 95 percent confidence intervals for these variables included or were very close to 1.0. Discussion The findings of this study demonstrated that the amount and type of milk consumed by mothers strongly predicted the amount and type of milk consumed by their school-aged children. This study also demonstrated that differences in children's milk consumption patterns were associated with a number of demographic variables. These included regional differences; differences associated with mothers' level of education; and children's age, gender, and race. Limitations The problem of underreporting of food intake is a concern when interpreting dietary intake data (24) . When food consumption surveys are used to obtain dietary intake data, both adults and adolescents tend to underreport their food intake (22). However, there is agreement that individuals of all ages are prone to exaggerate those foods they perceive to be healthful and to underreport foods that are commonly considered "sin" foods (i.e., foods high in sugar and fat) (22). Milk is generally perceived as a healthful food and was not among those foods most likely to be underreported in the CSFII (18). Hence underreporting was not likely to be a significant problem in this study. Factors Influencing the Amount of Milk Consumed by U.S. School-Aged Children The amount of milk consumed by mothers was associated strongly and positively with the amount of milk consumed by their children. Parents guide and direct children's food choices (17). Wardle and colleagues studied parental influences on children's consumption patterns and found significant mother-child correlations for consumption of dietary fat as well as fruit and vegetable consumption (34). Harper and Sanders observed that children sample unfamiliar food consistently more often when they view their parents partaking of the food (10). Children whose mothers do not drink milk may be less likely to sample milk, perceiving milk as an unfamiliar food. Parental monitoring may also influence children's milk consumption. Research has shown that parental monitoring can have a marked effect on children's food selection (I 7). Researchers interviewed over 50 focus groups with children nationwide regarding the factors that influence their consumption of calcium-rich foods. They discovered that a large percentage of children were neither encouraged nor required by their parents to drink milk at home (36). In our study, other predictors of the amount of milk consumed by U.S. school-aged children included children's gender and region. Compared with the girls, the boys consumed an average of I 28 grams per day more milk. This is an important finding, because girls' calcium intakes are also lower than boys' (4) . Girls' energy needs are typically lower than boys'. These lower energy needs may be reflected in lower intakes of all foods and beverages, including milk. On the Family Economics and Nutrition Review other hand, it is possible that some girls may be restricting their food intake . and eliminating or reducing thetr milk intake to cut calories and fat. Girls may initiate dieting behaviors as early as age 6. In one Ohio study of schoolchildren Grades 1 through 5, close to twice as many girls as boys reported restricting or altering their food intake (1). Adequate calcium int~ke is . especially important for grrls-bemg female is an independent risk factor for developing osteoporosis (8). In our study, differences found in children's milk intake by region of residence are also important. Southern girls have the lowest calcium int~kes, compared with girls in other regions (1 2). This study determined that children in the South also have the lowest milk intakes. In addition, they were more likely than children from other regions to drink no milk at all. Increased milk consumption among children in the South could be influential in improving their calcium intakes. Race did not predict the amount of milk consumed. Lactose maldigestion appears to vary widely among different ethnic and racial groups and in the United States is estimated to be about 15 percent in Whites, 80 percent in African Americans, and 90 percent in Asian Americans (19). However, a dairy-rich diet was found to be well tolerated when fed to AfricanAmerican adolescent girls for 21 days (29) . In this study race did not infl~ence total milk intake. This is consistent with findings that most people with lactose maldigestion are able to tolerate a glass of milk at a meal without developing any significant symptoms (32). 2001 Vol. 13 No. 1 Factors Influencing the Type of Milk Consumed by U.S. School-Aged Children The results of this study also demonstrated that a variety of factors influenced the type (skim, 2%, 1%, whole, none) of milk consumed by U.S. . school-aged children. The type of milk consumed by the mothers was associated strongly with the type of milk consumed by their children. This finding was consistent with results of studies conducted by Fischer and Birch, demonstrating that exposure to a food over time will result in the development of a preference for the food among children (6) . Children who have continued exposure to 1% and skim milk in the home and who observe their mothers consuming these types of milk are likely also to dri~ these types of milk. The type of milk consumed by children can have an effect on total diet quality. Children who drink skim milk come closer to meeting dietary recommendations f~r fat and saturated fat in their total daily diet (1 5,28) . Individuals in all age groups who consume I% and skim milk also consume more fruits and vegetables and less red meat (20). On the other hand, the cross-sectional nature of these data make it difficult to sort out the directionality of the association between the type of milk consumed by mothers and the type consumed by their children. Thus, it is possible that mothers m~y simp_ly drink the type of milk their children hke, and if the children do not like milk, mothers may not buy it just for themselves. Prior studies have shown that mothers' education level is correlated with their children's nutrient intake (13). In our study, mothers with the fewest years of education were more likely to have children who drank whole milk or no milk at all, compared with mothers who were more highly educated. Nutrition information may not be reaching less educated mothers. It is also possible that 1% and skim milk are not as accessible to them. Whole milk is sometimes the only choice available in lower income communities (35). Children from the South (compared with those in other regions) as well as black children (compared with white children) were more likely to drink whole milk or no milk at all. It may be necessary to target the Southern United States for outreach, because children in the South have the highest fat and saturated fat intakes and the lowest calcium intakes of children in all regions in the United States (14). Several other variables were associated with the type of milk consumed by school-aged children. Older children and girls were more likely to dri~ skim milk than were younger children and boys, respectively. Findings als~ showed that children eligible to receive free school lunch were less likely to drink skim milk than were children who paid full price. Beginning in the fall of 1996, schools participating in USDA school nutrition programs were required by law to serve meals that on average meet the dietary guidelines for fat, saturated fat, cholesterol, and sodium (5) . Because it is difficult to meet the dietary guidelines when a meal includes whole milk (1 5), participating schools may now be_ serving and marketing I% and skun milk more vigorously. Further research using future USDA surveys is needed to conftrm this possibility. 33 Implications The findings of our study demonstrate that mothers' milk consumption patterns are potentially strongly associated with the type and amount of milk consumed by U.S. school-aged children. Interventions aimed at increasing children's milk consumption should consider the strong influence of maternal modeling on children's milk intake. Mothers should be encouraged to serve as positive role models for their children by drinking skim or 1% milk regularly. In addition, it becomes apparent that milk promotion campaigns targeting women for prevention of osteoporosis may have a spillover effect of increasing children's milk consumption. Acknowledgments This project was funded by the Vermont Agricultural Experiment Station Hatch Project #VT-NS-00577. 34 References I. Berg, F.M. 1997. Afraid to Eat; Children and Teens in Weight Crisis (2nd ed.). Healthy Weight Publishing Network, Hettinger, North Dakota. 2. Cadogen, J., Eastell, R., and Jones, N. 1997. Milk intake and bone mineral acquisition in adolescent girls: Randomized, controlled intervention trial. BMJ 15: 1255-1260. 3. Chan, G.M., Hoffman, K., and McMurry, M. 1995. Effects of dairy products on bone and body composition in pubertal girls. Journal of Pediatrics 126:551-556. 4. Continuing Survey of Food lntakes by Individuals (CSFII), 1994-95 data set descriptions and data tables: Combined results of the USDA's 1994-95 CSFII. ARS Beltsville Human Nutrition Research Center, Food Surveys Research Group, l996. 5. Eadie, R.E. 1995. Child nutrition programs: School meal initiatives for healthy children; final rule. Federal Register 60(113):31188-31222. 6. Fischer, J.O. and Birch, L.L. 1995. Fat preferences and fat consumption of 3-5 year-old children are related to parental adiposity. Journal of the American Dietetic Association 95:759-764. 7. Fowler, F.J. 1990. Survey Research Methods. Applied Social Research Methods Series, Volume I. Sage Publications, Newbury Park, CA. 8. Guthrie, H.A. and Picciano, M.F. 1995. Human Nutrition. Mosby, Boston. 9. Harnack, L., Stang, J., and Story, M. 1999. Soft drink consumption among U.S. children and adolescents: Nutritional consequences. Journal of the American Dietetic Association 99:436-441 . 10. Harper, L.V. and Sanders, K.M. 1975. The effects of adults on young children's acceptance of unfamiliar foods. Journal of Experimental Child Psychology 20:206-214. II. Johnson, R.K., Driscoll, P., and Goran, M.I. 1996. Comparison of multiple-pass 24- hour recall estimates of energy intake with total energy expenditure determined by the doubly labeled water method in young children. Journal of the American Dietetic Association 96: 1140-1144. 12. Johnson, R.K., Guthrie, H., and Smiciklas-Wright, H. 1994. Characterizing nutrient intakes of children by sociodemographic factors. Public Health Reports I 09(3):414-420. 13. Johnson, R.K., Johnson, D., Harvey, J., and Wang, M. 1994. Dietary quality of the noon-time meal among a large sample of U.S. adolescents. School Food Service Research Review 18(1):2-7. 14. Johnson, R.K., Johnson, D., Wang, M., Smiciklas-Wright, H., and Guthrie, H. 1994. Characterizing nutrient intakes of adolescents by sociodemographic variables. Journal of Adolescent Health 15:149-154. 15. Johnson, R.K., Panely, C.V., and Wang, M.Q. 1998. The association between noontime beverage consumption and the diet quality of school-aged children. Journal of Child Nutrition and Management 22(2):95-1 00. 16. Johnston, C.C., Miller, J.Z., Slemenda, D.W., Reister, T.K., Hui, S., Christian, J.C., and Peacock, M. 1992. Calcium supplementation and increases in bone mineral density in children. The New England Journal of Medicine 327:82-87. Family Economics and Nutrition Review 17. Klesges, R.C., Stein, R.J., and Eck, L.H. 1991. Parental influence on food selection in young children and its relationship to childhood obesity. The American Journal of Clinical Nutrition 53:859-864. 18. Krebs-Smith, S.M., Graubard, B., Cleveland, L., Subar, A., Ballard-Barbash, R., and Kahle, L. 2000. Low energy reporters vs others: A comparison of reported food intakes. European Journal of Clinical Nutrition 54:281 -287. 19. Lactose Intolerance. 1994. National Digestive Diseases Information Clearinghouse, Washington, DC. National Institutes of Health publication 94-2751. 20. Lee, H.C., Gerrior, S.A., and Smith, J.A. 1998. Energy, macronutrient and food intakes in relation to energy compensation in consumers who drink different types of milk. The American Journal of Clinical Nutrition 67:616-623. 21. Lee, W., Leung, S., and Wang, S.H. 1994. Double-blind, controlled calcium supplementation and bone mineral accretion in children accustomed to a low calcium diet. The American Journal of Clinical Nutrition 60:744-750. 22. Livingstone, M.B., Prentice, A.M., and Coward, W.A. 1992. Validation of estimates of energy intake by weighted dietary record and diet history in children and adolescents. The American Journal of Clinical Nutrition 56:29-35. 23. Lloyd, T., Andon, M.B., and Rollings, N. 1993. Calcium supplementation and bone mineral density in adolescent girls. lAMA 270(7):841-844. 24. Mertz, W., Tsui, J.C., and Judd, J.T. 1991 . What are people really eating? The relation between energy intake derived from estimated diet records and intake determined to maintain body weight. The American Journal of Clinical Nutrition 54:28-35. 25. Morel, J.G. 1989. Logistic regression under complex survey designs. Survey Methodology 15:203-223. 26. Oliveria, S.A., Ellison, R.C., and Moore, L.L. 1992. Parent-child relationships in nutrient intake: The Framingham Children 's Study. The American Journal of Clinical Nutrition 56:593-598. 27. Pelletier, D.L., Kendall, A., and Mathios, A. 1996. Lowfat milk promotion: Opportunities created by a new policy environment. Unpublished work, pp. 1-25. Cornell University, Ithaca, NY. 28. Peterson, S. and Sigman-Grant, M. 1997. Impact of adopting lower-fat food choices on nutrient intake of American children. Pediatrics I 00(3):A380. 29. Pribila, B., Hertzler, S.R., Martin, B.R., Weaver, C.M., and Savaiano, D.A. 2000. Improved lactose digestion and intolerance among African-American adolescent girls fed a dairy-rich diet. Journal of the American Dietetic Association I 00:524-528. 30. Sentipal, J.M., Wardlaw, G.M., and Mahan, J. 1991. Influence of calcium intake and growth indexes on vertebral bone mineral density in young females. The American Journal of Clinical Nutrition 54:425-428. 31 . Shah, B.V., Barnwell, B.G., and Bieler, G.S. 1997. SUDAAN Software for the Statistical Analysis of Correlated Data. User's Manual. Research Triangle Institute, NC. 32. Suarez, F.L., Savaiano, D.A., and Levitt, M.D. 1995. A comparison of symptoms after the consumption of milk or lactose-hydrolyzed milk by people with self-reported severe lactose intolerance. The New England Journal of Medicine 333:1-4. 2001 Vol. 13 No. I 35 33. Thompson, F.E. and Dennison, B.A. 1994. Dietary sources of fats and cholesterol in U.S. children aged 2 through 5 years. American Journal of Public Health 84:779-806. 34. Wardle, J., Gibson, L., and Watts, C. 1995. Parental influences on children's nutritional knowledge, diet and risk factor status. Annals of Behavioral Medicine PA16A, S079. 35. Wechsler, H., Basch, C.E., Zybert, P., Lantigua, R., and Shea, S. 1995. The availability of low-fat milk in an inner-city Latino community: Implications for nutrition education. American Journal of Public Health 85(12) : 1690-1692. 36. Wolfe, F.H. Factors influencing the intake of calcium rich foods among adolescents. [On-line). Available: http://cristel.nal.usda.regional.linkpage. Accessed December 13, 1999. 36 Family Economics and Nutrition Review Carol Byrd-Bredbenner, PhD, RD Rutgers, The State University of New Jersey Darlene Grasso, MA, RD Montclair State University 2001 Vol. 13 No. l The Effects of Food Advertising Policy on Televised Nutrient Content Claims and Health Claims This study examined changes in nutrient content and health claims made in televised food advertisements before and after the Federal Trade Commission's 1994 food advertising pol icy, which is predicated on the Nutrition Labeling and Education Act (NLEA). Our sample included 1 05 and 1 08 advertisements broadcast during prime-time in 1992 and 1998, respectively. The rate that nutrient content and health claims were used was low in both years. And none of the advertisements contained diet-disease health claims authorized by the Food and Drug Administration. Although current food advertising policy virtually eliminates deceptive advertisements, it may also limit diet-disease health claims in broadcast media . More flexibility in presenting diet-disease health cla ims in broadcast media advertising could increase the use of such claims and contribute to the goal of NLEA to educate consumers. T he decision to purchase a food is influenced by many factors, one of which is advertising (7,8,36,43). Advertisements traditionally promoted foods and beverages by featuring mainly sensory qualities, convenience, and economic factors (10,44). In recent years some of these advertisements have tried to influence consumer-purchasing decisions by also touting nutritional or health qualities or both (29,32). Food advertising, like advertising for nearly all products, is regulated by the Federal Trade Commission (FTC). Historically, the FTC permitted nutrient claims (e.g., "high in fiber") in advertising and never formally prohibited diet-disease health claims (i.e., claims that explicitly linked the consumption [or lack of consumption] of a particular nutrient or other substance in a food to a disease or health-related condition [e.g., "a calcium-rich diet can help prevent osteoporosis"]) (32) . However, if diet-disease health claims were made on the label, the Food and Drug Administration (FDA) reclassified the food as a drug and required the manufacturer to adhere to the drugapproval procedures of the FDA (29). For years food advertisers did not make diet-disease health claims about their products, but as the connection between diet and health became increasingly clear, food manufacturers and advertisers grew interested in using this information to sell their products. Consequently, in 1984, the Kellogg Company initiated an advertising campaign that explicitly described the relationship between a high-fiber diet and reduced risk of certain types of cancer. When the FDA failed to prosecute this direct violation of 37 diet-disease health claims, other food manufacturers launched similar campaigns (18,29). Marketing strategies that included diet-disease health claims did provide consumers with information about nutrition and health. However, in their zeal to gain a competitive edge, advertisers also pushed the limits of what science could support and what consumers would believe (18,25). To stem questionable marketing practices and restore consumer confidence, the Nutrition Labeling and Education Act (NLEA) was passed in 1990 and became fully effective in 1994 (27). The NLEA overhauled nutrition labels on food packages, expanded the scope of nutrition labeling, explicitly defined nutrient content claims, and regulated dietdisease health claims (25). While the new food-labeling regulations did much to improve the quality of information on food packages, these regulations did not extend to food advertising (41). Fortunately, in its efforts to prevent deceptive or misleading claims, the FTC announced in 1994 that it would apply the standards set forth in the NLEA to evaluate nutrient content and diet-disease health claims made in food advertisements (14). The FTC reported that its goal was to create a food advertising policy that would help ensure that food advertising messages are congruent with data presented and are permitted on food labels (1 5). While food and beverage advertisements appear in all types of print, broadcast, and electronic media, television is the preferred advertising medium of food manufacturers--over 7 5 percent of their 1997 advertising budget was spent on televised advertising (17). The food and alcohol industry accounted for more than one-sixth of the $73-billion mass media advertising market; only the automobile industry spent more on advertising ( 17). 38 Although some studies have examined the nutrient content claims and health claims in food advertising, few have focused on televised advertising. Furthermore, no studies could be located that compared changes in nutrient content claims and health claims over time or examined the effect of the NLEA and FTC food advertising policy on televised food advertisements. Thus the purpose of this study was to examine changes in the nutrient content claims and health claims made in televised food advertisements before and after the enactment of the new food advertising policy of the FTC, a policy which is based on the NLEA, and to determine whether the use of claims varied by type of food product advertised. Methods Sample In the autumn ofboth 1992 and 1998, 17.5 hours of top-ranked, prime-time 1 were videotaped. This study focused on prime-time and major networks because they traditionally have the largest viewing audience (35). The sample comprised all commercials broadcast during the sampling period. Commercials (i.e., all non-program time) included advertisements, public service announcements, and promotions for television programs. Although all commercials were recorded and analyzed, only data pertaining to food advertisements are presented here. A food advertisement was defined as a paid-commercial announc;ement that specifically promoted a food, beverage, or dietary supplement intended for human consumption. 1 Prime-time refers to programming broadcast from 8 p.m. to 11 p.m. Monday through Saturday, and 7 p.m. to II p.m. on Sunday. Major networks refer to ABC, CBS, NBC, Fox, and WB; note WB became a network in 1998. Instrument The food advertisements were content and textually analyzed by using the study instrument that was adapted from those reported elsewhere (5,19,28,38, 40,50). Content analysis permits systematic, objective evaluation of visual and linguistic elements (6,24) . Textual analysis allows researchers to investigate how linguistic elements are used, their significance, and their contribution to understanding a topical area (4,38). Content analysis began by eliminating all nonfood commercials. All food advertisements were then classified into 11 food categories based largely on the USDA Food Guide Pyramid (47): Breads and cereals, vegetables, fruits, protein-rich foods (i.e., eggs, meat, poultry, fish, shellfish, nuts, and seeds), dairy products, high-sugar foods (e.g., syrup, candy, and soft drinks), high-fat foods (e.g., butter, oils, and salad dressing), alcohol-containing beverages (i.e., wine), calorie-free beverages, dietary supplements, and miscellaneous items (i.e., seasonings). Restaurant advertisements frequently highlighted a variety of food items that together comprised a meal. Thus to evaluate the nutritional value of the foods advertised, we assigned all items in an advertised meal to the appropriate food categories. In addition, combination foods (e.g., fast-food sandwiches and soups) were broken down into their component parts and appropriately assigned to two or more of the food categories. Foods in the first five categories listed previously were further classified by nutrient density: low, moderate, and high. Methods described in detail elsewhere were used to classify density {51). In brief, foods low in nutrient density tended to be ones that are highest in fat in each of the first five categories (e.g., pastries, French fries, coconut, luncheon meats, and whole milk). Foods moderate in Family Economics and Nutrition Review nutrient density were less nutrient dense than were foods high in nutrient density (e.g., breads made with enriched flour instead of whole grains, candied sweet potatoes instead of plain vegetables, fruits canned in syrup rather than fresh or canned in unsweetened juice, fattrimmed beef instead of skinless poultry white meat, or lowfat instead of nonfat milk). Foods high in nutrient density provided the greatest level of nutrients per kilocalorie. The subsequent step, requiring textual analysis, involved identifying and coding nutrient content claims as either (a) Contains Specific Nutrient or (b) Minimizes (or eliminates) Specific Nutrient. Nutrient content claims, defined in the FDA and USDA's foodlabeling regulations, include 11 core terms that can be used to describe the nutrient content of foods: good source, more, high, free, low, lean, extra lean, reduced, less, light, and fewer (42). An advertisement that indicated |
OCLC number | 888048794 |
|
|
|
A |
|
C |
|
G |
|
H |
|
I |
|
N |
|
P |
|
U |
|
W |
|
|
|