- Email: [email protected]
Salad Bars and Fruit and Vegetable Consumption in Elementary Schools: A Plate Waste Study
RESEARCH Research and Professional Briefs
Salad Bars and Fruit and Vegetable Consumption in Elementary Schools: A Plate Waste Study MARC A. ADAMS, MPH; ROBIN L. PELLETIER, MPH; MICHELLE M. ZIVE, MS, RD; JAMES F. SALLIS, PhD
ABSTRACT The object of this study was to determine whether students attending schools with self-service salad bars consume a greater amount of fruits and vegetables compared with students using preportioned servings and to evaluate the relationship between number of items offered and fruit and vegetable consumption. Two hundred ninetyfour students in first through fifth grade were randomly selected from two schools with salad bars and two with preportioned servings. Weights of fruit and vegetable items were measured pre- and postconsumption and interobserver agreement ⫾1 g was ⱖ95%. Presence of a salad bar was not associated with greater fruit and vegetable consumption. Fruit and vegetable consumption was positively related to the number of fruit and vegetable items offered at salad bars (P⬍.05), adjusting for sex and grade. Fruit and vegetable variety was associated with elementary school–age children’s fruit and vegetable consumption when using salad bars. J Am Diet Assoc. 2005;105:1789-1792.
F
ruit and vegetable consumption has been linked to the prevention of chronic diseases such as diabetes, some cancers, heart disease, and obesity (1). Health organizations such as the American Dietetic Association strongly advise people of all ages to eat at least five servings of fruits and vegetables every day (2,3). American children only eat an average of three and a half servings daily (1,4). Methods to increase school-aged children’s fruit and vegetable consumption have emphasized classroom curricula (5). However, school environment interventions combined with classroom curricula have shown greater promise for changing behaviors (6,7).
M. A. Adams is evaluation coordinator, R. L. Pelletier is program director, M. M. Zive is a registered dietitian, and J. F. Sallis is principal investigator, Graduate School of Public Health, San Diego State University, San Diego, CA. Address correspondence to: Marc A. Adams, MPH, Graduate School of Public Health, San Diego State University, 3900 Fifth Ave, Suite 310, San Diego, CA 92103. E-mail: [email protected] Copyright © 2005 by the American Dietetic Association. 0002-8223/05/10511-0009$30.00/0 doi: 10.1016/j.jada.2005.08.013
© 2005 by the American Dietetic Association
Because most children eat at least one meal at school per school day, the school cafeteria remains an important setting for exploring connections between the environment and students’ dietary habits. The National School Lunch Program feeds about 27 million students daily (8) and provides opportunities to increase students’ fruit and vegetable consumption (9). The US Department of Agriculture report, School Lunch Salad Bars, states that a wide range of fruits and vegetables are available in salad bars and, during the past 10 years, prevalence of salad bars increased in elementary schools (10). Although implementing salad bars has been a strategy to promote fruit and vegetable consumption among school-aged children (10-13), empirical evidence of increased consumption is lacking. Accurate measurement of children’s fruit and vegetable consumption remains a challenge. The most common method for dietary assessment is the use of 24-hour recalls, but problems with self-reports are well documented (14). Observations are considered a better alternative, but observers overestimate fruit and vegetable consumption (15) and changes are difficult to detect (16). The most precise method for dietary assessment may be measuring the pre and post weight of a participant’s plate. However, this time-consuming method is rarely used (16). Salad bars have become a recommended method to increase fruit and vegetable consumption without strong evidence of their effectiveness. The purpose of our study was to use plate waste weighing to determine if elementary school students with self-service salad bars consume more fruits and vegetables compared with students served preportioned amounts determined by cafeteria staff. METHODS Before the study, the Institutional Review Board at San Diego State University and principals from each school approved the study. Four elementary schools within two San Diego County districts were selected. The schools were predominantly nonwhite with Latino/Hispanics making up the greatest proportion of minority students (Table 1). Most students qualified for free or reducedprice lunches. One district used daily salad bars and the other used preportioned fruit and vegetable servings. The daily salad bars were a permanent feature in the cafeterias, existing at least 10 years before the study. The study was conducted 1 day at each school during April or May 2003. One salad bar and one preportioned school were
Journal of the AMERICAN DIETETIC ASSOCIATION
1789
Table 1. Demographic profile of first through fifth graders in four California schools and lunch items served by school and serving type Serving type
Non-white (%)
Free or reduced lunch (%)
Types of fruit and vegetable and entrées served
District 1 Preportioned school 1
67
52
Preportioned school 2
79
77
Green beans, carrot sticks, canned pears, whole apples, and whole oranges (chicken nuggets and tater totsa) Caesar salad, green beans, carrot sticks, pear quarters, and whole apples (chicken fajitasa)
District 2 Salad bar school 1
65
89
Salad bar school 2
45
64
a
Caesar salad, green beans, canned pears, and whole apples (chicken dino bitesa) Caesar salad, corn, cucumber slices, carrot sticks, lime wedges, whole apples, and whole oranges (mini corn dogs with macaroni and cheesea)
Entrée items are displayed in parentheses.
sampled on a Monday and the other two schools on Friday to control for any day-of-week effects. Using a cross-sectional design, 294 first through fifth graders were randomly selected for plate waste measures. Estimates of sample size for this study were calculated from the effects observed by Kirks and Wolff (16), who conducted a plate waste study among kindergarten through sixth-grade students. Six university-level research assistants helped collect data. Research assistants underwent 10 hours of training on student selection and measurement before working with students. Students whose school lunch identification numbers matched a list of random numbers were invited to have their aggregate fruit and vegetable items weighed using gram scales. Those who assented to the study were given a tray with colored stickers attached to identify them as participants. A small gift, such as a miniature plastic ruler or kaleidoscope, was promised if they completed both pre- and postweighing. Salad bar and preportioned schools had similar setups for distributing lunches. Each elementary school had a 30-minute lunch period, with students going through the lunch line at the beginning of the period. All classes are served within the first 15 minutes of the lunch period. Students first selected their fruit and vegetable items and then the entrée. The salad bar setups were also similar, in that students accessed the salad bar items on both sides of the standardized equipment and formed one line to proceed to the entrée serving station. Students with a salad bar placed fruit and vegetable items directly on their trays and could select any items and amounts. Preportioned items were placed by cafeteria staff into standardized paper cups or plastic bags for students to choose. For preportioned items, actual amounts differed because cafeteria staff determined the servings. Table 1 presents items offered at each school. Fruits and vegetables were measured to the nearest gram after students selected their fruit and vegetable items but before the entrée items. Any salad, fruit, vegetable, or dressing item, not part of the entrée or dessert, was included in the collective weight. A second measurement was conducted as students disposed of their trays. Remaining fruit and vegetable items were scraped onto paper plates for individual postlunch weighing. Weights
1790
November 2005 Volume 105 Number 11
of the serving tray, cups, bags, measurement cups, and paper plates were subtracted from each measurement to produce a net weight. A 10% sample of students at each measurement point had trays reweighed by a second observer. Interobserver agreement ⫾1 g was ⱖ95% throughout the study. Students’ sex and class level were transcribed by identification numbers associated with school records. For missing sex information, two raters independently assigned a sex based on the student’s first name. Disagreements were resolved by consensus. Consumption was calculated by subtracting the fruit and vegetable postconsumption weight from preconsumption weight. Fruit and vegetable variety was defined as the number of fruit and vegetable items served that were not part of the entrée or dessert. Mean weights were computed for each school serving type (ie, salad bar or preportioned). The Statistical Analysis System (version 8.01, 2000, SAS Institute, Inc, Cary, NC) was used to calculate means and standard deviations for all analyses. An analysis of covariance determined if the presence of a salad bar or variety affected the amount taken or consumed while controlling for sex and grade level. Differences in adjusted means were evaluated by Tukey post hoc comparisons. The proportion of variation in the dependent variable explained by the independent variables (2) is also presented. RESULTS AND DISCUSSION Of the 294 students who assented, complete consumption data were available for 288 students (98%). The final sample was 54% girls. The participant distribution across grade levels consisted of 20% first graders, 20% second graders, 23% third graders, 23% fourth graders, and 15% fifth graders. Students who attended schools with salad bars present took 112⫾70 g fruits and vegetables on average compared with 104⫾86 g taken by students at non–salad bar schools. These differences were not statistically significant and the variance accounted for was small (2⫽0.06). The mean amounts consumed were also not significantly different between serving type (47⫾60 g vs 43⫾58 g, respectively). Analysis of covariance revealed that consumption was significantly different in schools with fruit
Table 2. Adjusted mean weights of fruits and vegetables taken and consumed by first- through fifth-grade students in four California schools (n⫽288) Serving type District 1 Preportioned school 1 Preportioned school 2 District 2 Salad bar school 1 Salad bar school 2 a
No. of students
Items served (n)
Items taken (g)
Items consumed (g)
Items consumed (%)
a
4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ mean⫾SE ™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 76 5 96⫾10 49⫾7 51 65 5 96⫾12 38⫾8 40 70 77
4 7
117⫾10 107⫾10
36⫾7 61⫾7
31 57
SE⫽standard error.
and vegetable variety (F⫽2.83, Pⱕ.05). The number of items offered varied by school with the two salad bar schools offering four and seven items each, and the preportioned schools offering five items each. Tukey post hoc analysis suggested a difference in consumption between students in the highest and lowest variety schools (P⫽.06), with a small effect size (2⫽0.08). A trend in increasing consumption was noticed as variety increased. Fruit and vegetable weights adjusted for sex and grade levels are presented in Table 2. Salad bars are common in public elementary schools and have become more prevalent in the past 10 years (10). To date no study has reported how the presence of salad bars or fruit and vegetable variety is related to consumption in elementary school students. Our results do not support the hypothesis that salad bars increase fruit and vegetable consumption among elementary school students. We found that students consumed on average 4 g more (⬍3% of a medium apple) fruits and vegetables with a salad bar than students using preportioned servings. These differences are not statistically or practically significant, and are inconsistent with the limited body of salad bar research in schools (17,18). The previous studies were based on either self-report or observational methods. Children’s ability to recall intake is prone to misreporting (19), and observers have been found to overestimate consumption (15,20). Therefore, actual measurements of combined fruit and vegetable weights were used in this study. Measuring the weight of food on each plate produces more precise consumption results than other methods (16,21). We found fruit and vegetable variety of salad bars was associated with greater consumption. Item variety is usually higher with salad bars (10). However, the relation of salad bar variety to consumption has not been evaluated. Schools with the preportioned items offered five items each. Salad bar schools offered four or seven items. The salad bar school with the greatest variety of items offered also had the highest mean amounts of fruits and vegetables consumed, and the school with the lowest variety of items offered had the lowest consumption. These findings suggest that fruit and vegetable variety may mediate the effect of salad bars on fruit and vegetable consumption. Our results provide some suggestions about the mechanisms by which fruit and vegetable variety could increase consumption. Table 2 shows the amount of fruits and vegetables taken was lower in the high-variety salad
bar school compared with the low-variety school, but that the amount consumed at the high-variety school was greater. Perhaps more variety made it more likely that children would find preferred fruits and vegetables. It is surprising that students took larger servings from the low-variety salad bar, but wasted more. This study had several limitations. Because larger differences between serving methods and smaller standard deviations were expected, the power to detect significant differences was limited. A small number of schools were studied, and each school was sampled 1 day only. To control for specific preference of fruits and vegetables and day of the week effects across serving methods, schools were sampled on either a Monday or Friday when they offered similar fruit and vegetable items. Consumption was higher on Mondays compared with Fridays, and higher variety schools were sampled on Mondays and low variety schools on Fridays. Thus, variety and day of the week effects may not be independent. Future studies may want to sample schools on 1 day only. This approach would limit generalizability to other weekdays. An alternative would be to study all schools on all days of the week, which would be preferable and allow for true testing of day-of-week effects. The difference in consumption between salad bars could be due to fruit and vegetable item variety, day of the week, or factors not measured, such as the location of the salad bar, presentation quality, popularity of specific fruits and vegetables, signage, student demographics, or social norms within the school. It should be noted that the school with the highest consumption also had the lowest percent of nonwhite students, so future studies should control for student ethnicity and socioeconomic status. It is possible that present results are not generalizable to other schools. Our results provide initial evidence that fruit and vegetable variety is related to elementary children’s fruit and vegetable consumption. The number of items offered may mediate the effectiveness of salad bars. Future studies should determine whether fruit and vegetable variety affects consumption with both salad bars and preportioned servings while controlling for students’ demographic and economic characteristics. Other potential environmental mediators of salad bar effectiveness, such as location, presentation, and interaction with entrees, also need to be examined.
November 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1791
CONCLUSIONS Environmental interventions combined with educational approaches may be needed to effectively promote children’s fruit and vegetable consumption at school. This study suggests that merely providing a self-service salad bar is not a sufficient intervention to improve elementary school children’s fruit and vegetable intake. However, salad bars with more variety were associated with higher intakes. If confirmed in further studies, salad bars with a high variety of fruits and vegetables may become a recommended strategy for meeting national fruit and vegetable intake goals. This research was funded by a US Department of Agriculture Food Stamp Program award to James F. Sallis, PhD, through the California Nutrition Network (Contract 03-75578). The authors thank the Cajon Valley Union and Lemon Grove School Districts; Linda Patzold, RD, and Patricia McKay, directors of Child Nutrition Services; the cafeteria staff at Magnolia, Ballantyne, Johnson, Mount Vernon, and Vista La Mesa schools; project coordinator Justine Kozo; and research assistants Ayodeji Oladunjoye, Veronica Guzman, Lisa Hopper, and Gary Gailius for their help with this project. References 1. Fruits and Vegetables. Enhanced Federal Efforts to Increase Consumption Could Yield Health Benefits for Americans. Washington, DC: US General Accounting Office; 2002. Publication GAO-02-657. 2. Cummings S, Parham ES, Strain GW. Position of the American Dietetic Association: Weight management. J Am Diet Assoc. 2002;102:1145-1155. 3. US Dept of Agriculture. Nutrition and Your Health: Dietary Guidelines for Americans. 5th ed. Available at: http://www.health.gov/dietaryguidelines. Accessed April 9, 2004. 4. Krebs-Smith SM, Cook A, Subar AF, Cleveland L, Friday J, Kahle LL. Fruit and vegetable intakes of children and adolescents in the United States. Arch Pediatr Adolesc Med. 1996;150:81-86. 5. McGraw SA, Sellers D, Stone E, Resnicow KA, Kuester S, Fridinger F, Wechsler H. Measuring implementation of school programs and policies to promote healthy eating and physical activity among youth. Prev Med. 2000;31(suppl):S86-S97. 6. Weschler H, Devereaux R, Davis M, Collins J. Using the school environment to promote physical activity and healthy eating. Prev Med. 2000;31(suppl):S121S137. 7. Birnhaum AS, Lytle LA, Story M, Perry CL, Murray DM. Are differences in exposure to a multicomponent school-based intervention associated with varying dietary outcomes in adolescents? Health Educ Behav. 2002;29:427-443. 8. Briggs M, Safaii S, Beall DL. Position of the Ameri-
1792
November 2005 Volume 105 Number 11
9.
10.
11.
12.
13.
14.
15.
16. 17.
18.
19. 20. 21.
can Dietetic Association, Society for Nutrition Education, and American School Food Service Association: Nutrition services: An essential component of comprehensive school health programs. J Am Diet Assoc. 2003;103:505-514. Buzby J, Guthrie J. Food Assistance Research Brief— The USDA Fruit and Vegetable Pilot Program Evaluation. Washington, DC: US Dept of Agriculture; 2003. Food Assistance and Nutrition Research Report No. 34-14. US Dept of Agriculture, Food and Nutrition Service, Office of Analysis, Nutrition and Evaluation. School Lunch Salad Bars. Alexandria, VA: US Department of Agriculture; 2003. School Lunch Program: Efforts Needed to Improve Nutrition and Encourage Healthy Eating. Washington, DC: US General Accounting Office; 2003. Publication GAO 03-506. Buzby JC, Guthrie JF. Plate waste in school nutrition programs: Final report to Congress. Economic Research Service E-FAN-02-009. Available at: www. ers.usda.gov/publications/efan02009/. Accessed March 14, 2004. US Dept of Agriculture. The School Nutrition Dietary Assessment Study: School Food Service, Meals Offered and Dietary Intakes. Princeton, NJ: Mathematica Policy Research Inc; 1993. MPR Reference No. 7937-140. Buzzard MI, Faucett CL, Jeffrey RW, McBane L. Monitoring dietary change in a low-fat diet intervention study: Advantages of using 24-hour dietary recalls vs food records. JAMA. 1996;96:574-579. Shankar AV, Gittelsohn J, Stallings R, West KP Jr, Gnywali T, Dhungel C, Dahal B. Comparison of visual estimates of children’s portion sizes under both shared-plate and individual-plate conditions. J Am Diet Assoc. 2001;101:47-52. Kirks BA, Wolff HK. A comparison of methods for plate waste determinations. J Am Diet Assoc. 1985; 85:328-331. Slusser W, Neumann C. Evaluation of the effectiveness of the salad bar program in the Los Angeles Unified School District. Available at: http://socialmarketingnutrition.ucdavis.edu/VIDEO/video2002.htm. Accessed March 14, 2003. Kerfoot B, Fournet R. Evaluation of the use of a self-service fruit and vegetable bar on the consumption and plate waste. J Am Diet Assoc. 1996;96(suppl 9):A-10. Livingstone MB, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc. 2000;59:279-293. Gittelsohn J, Shankar AV, Pokhrel RP, West KP Jr. Accuracy of estimating food intake by observation. J Am Diet Assoc. 1994;94:1273-1277. Comstock EM, Syminton LE. Distributions of serving sizes and plate waste in school lunches. Implications for measurement. J Am Diet Assoc. 1982;81:413-422.
Salad Bars and Fruit and Vegetable Consumption in Elementary Schools: A Plate Waste Study MARC A. ADAMS, MPH; ROBIN L. PELLETIER, MPH; MICHELLE M. ZIVE, MS, RD; JAMES F. SALLIS, PhD
ABSTRACT The object of this study was to determine whether students attending schools with self-service salad bars consume a greater amount of fruits and vegetables compared with students using preportioned servings and to evaluate the relationship between number of items offered and fruit and vegetable consumption. Two hundred ninetyfour students in first through fifth grade were randomly selected from two schools with salad bars and two with preportioned servings. Weights of fruit and vegetable items were measured pre- and postconsumption and interobserver agreement ⫾1 g was ⱖ95%. Presence of a salad bar was not associated with greater fruit and vegetable consumption. Fruit and vegetable consumption was positively related to the number of fruit and vegetable items offered at salad bars (P⬍.05), adjusting for sex and grade. Fruit and vegetable variety was associated with elementary school–age children’s fruit and vegetable consumption when using salad bars. J Am Diet Assoc. 2005;105:1789-1792.
F
ruit and vegetable consumption has been linked to the prevention of chronic diseases such as diabetes, some cancers, heart disease, and obesity (1). Health organizations such as the American Dietetic Association strongly advise people of all ages to eat at least five servings of fruits and vegetables every day (2,3). American children only eat an average of three and a half servings daily (1,4). Methods to increase school-aged children’s fruit and vegetable consumption have emphasized classroom curricula (5). However, school environment interventions combined with classroom curricula have shown greater promise for changing behaviors (6,7).
M. A. Adams is evaluation coordinator, R. L. Pelletier is program director, M. M. Zive is a registered dietitian, and J. F. Sallis is principal investigator, Graduate School of Public Health, San Diego State University, San Diego, CA. Address correspondence to: Marc A. Adams, MPH, Graduate School of Public Health, San Diego State University, 3900 Fifth Ave, Suite 310, San Diego, CA 92103. E-mail: [email protected] Copyright © 2005 by the American Dietetic Association. 0002-8223/05/10511-0009$30.00/0 doi: 10.1016/j.jada.2005.08.013
© 2005 by the American Dietetic Association
Because most children eat at least one meal at school per school day, the school cafeteria remains an important setting for exploring connections between the environment and students’ dietary habits. The National School Lunch Program feeds about 27 million students daily (8) and provides opportunities to increase students’ fruit and vegetable consumption (9). The US Department of Agriculture report, School Lunch Salad Bars, states that a wide range of fruits and vegetables are available in salad bars and, during the past 10 years, prevalence of salad bars increased in elementary schools (10). Although implementing salad bars has been a strategy to promote fruit and vegetable consumption among school-aged children (10-13), empirical evidence of increased consumption is lacking. Accurate measurement of children’s fruit and vegetable consumption remains a challenge. The most common method for dietary assessment is the use of 24-hour recalls, but problems with self-reports are well documented (14). Observations are considered a better alternative, but observers overestimate fruit and vegetable consumption (15) and changes are difficult to detect (16). The most precise method for dietary assessment may be measuring the pre and post weight of a participant’s plate. However, this time-consuming method is rarely used (16). Salad bars have become a recommended method to increase fruit and vegetable consumption without strong evidence of their effectiveness. The purpose of our study was to use plate waste weighing to determine if elementary school students with self-service salad bars consume more fruits and vegetables compared with students served preportioned amounts determined by cafeteria staff. METHODS Before the study, the Institutional Review Board at San Diego State University and principals from each school approved the study. Four elementary schools within two San Diego County districts were selected. The schools were predominantly nonwhite with Latino/Hispanics making up the greatest proportion of minority students (Table 1). Most students qualified for free or reducedprice lunches. One district used daily salad bars and the other used preportioned fruit and vegetable servings. The daily salad bars were a permanent feature in the cafeterias, existing at least 10 years before the study. The study was conducted 1 day at each school during April or May 2003. One salad bar and one preportioned school were
Journal of the AMERICAN DIETETIC ASSOCIATION
1789
Table 1. Demographic profile of first through fifth graders in four California schools and lunch items served by school and serving type Serving type
Non-white (%)
Free or reduced lunch (%)
Types of fruit and vegetable and entrées served
District 1 Preportioned school 1
67
52
Preportioned school 2
79
77
Green beans, carrot sticks, canned pears, whole apples, and whole oranges (chicken nuggets and tater totsa) Caesar salad, green beans, carrot sticks, pear quarters, and whole apples (chicken fajitasa)
District 2 Salad bar school 1
65
89
Salad bar school 2
45
64
a
Caesar salad, green beans, canned pears, and whole apples (chicken dino bitesa) Caesar salad, corn, cucumber slices, carrot sticks, lime wedges, whole apples, and whole oranges (mini corn dogs with macaroni and cheesea)
Entrée items are displayed in parentheses.
sampled on a Monday and the other two schools on Friday to control for any day-of-week effects. Using a cross-sectional design, 294 first through fifth graders were randomly selected for plate waste measures. Estimates of sample size for this study were calculated from the effects observed by Kirks and Wolff (16), who conducted a plate waste study among kindergarten through sixth-grade students. Six university-level research assistants helped collect data. Research assistants underwent 10 hours of training on student selection and measurement before working with students. Students whose school lunch identification numbers matched a list of random numbers were invited to have their aggregate fruit and vegetable items weighed using gram scales. Those who assented to the study were given a tray with colored stickers attached to identify them as participants. A small gift, such as a miniature plastic ruler or kaleidoscope, was promised if they completed both pre- and postweighing. Salad bar and preportioned schools had similar setups for distributing lunches. Each elementary school had a 30-minute lunch period, with students going through the lunch line at the beginning of the period. All classes are served within the first 15 minutes of the lunch period. Students first selected their fruit and vegetable items and then the entrée. The salad bar setups were also similar, in that students accessed the salad bar items on both sides of the standardized equipment and formed one line to proceed to the entrée serving station. Students with a salad bar placed fruit and vegetable items directly on their trays and could select any items and amounts. Preportioned items were placed by cafeteria staff into standardized paper cups or plastic bags for students to choose. For preportioned items, actual amounts differed because cafeteria staff determined the servings. Table 1 presents items offered at each school. Fruits and vegetables were measured to the nearest gram after students selected their fruit and vegetable items but before the entrée items. Any salad, fruit, vegetable, or dressing item, not part of the entrée or dessert, was included in the collective weight. A second measurement was conducted as students disposed of their trays. Remaining fruit and vegetable items were scraped onto paper plates for individual postlunch weighing. Weights
1790
November 2005 Volume 105 Number 11
of the serving tray, cups, bags, measurement cups, and paper plates were subtracted from each measurement to produce a net weight. A 10% sample of students at each measurement point had trays reweighed by a second observer. Interobserver agreement ⫾1 g was ⱖ95% throughout the study. Students’ sex and class level were transcribed by identification numbers associated with school records. For missing sex information, two raters independently assigned a sex based on the student’s first name. Disagreements were resolved by consensus. Consumption was calculated by subtracting the fruit and vegetable postconsumption weight from preconsumption weight. Fruit and vegetable variety was defined as the number of fruit and vegetable items served that were not part of the entrée or dessert. Mean weights were computed for each school serving type (ie, salad bar or preportioned). The Statistical Analysis System (version 8.01, 2000, SAS Institute, Inc, Cary, NC) was used to calculate means and standard deviations for all analyses. An analysis of covariance determined if the presence of a salad bar or variety affected the amount taken or consumed while controlling for sex and grade level. Differences in adjusted means were evaluated by Tukey post hoc comparisons. The proportion of variation in the dependent variable explained by the independent variables (2) is also presented. RESULTS AND DISCUSSION Of the 294 students who assented, complete consumption data were available for 288 students (98%). The final sample was 54% girls. The participant distribution across grade levels consisted of 20% first graders, 20% second graders, 23% third graders, 23% fourth graders, and 15% fifth graders. Students who attended schools with salad bars present took 112⫾70 g fruits and vegetables on average compared with 104⫾86 g taken by students at non–salad bar schools. These differences were not statistically significant and the variance accounted for was small (2⫽0.06). The mean amounts consumed were also not significantly different between serving type (47⫾60 g vs 43⫾58 g, respectively). Analysis of covariance revealed that consumption was significantly different in schools with fruit
Table 2. Adjusted mean weights of fruits and vegetables taken and consumed by first- through fifth-grade students in four California schools (n⫽288) Serving type District 1 Preportioned school 1 Preportioned school 2 District 2 Salad bar school 1 Salad bar school 2 a
No. of students
Items served (n)
Items taken (g)
Items consumed (g)
Items consumed (%)
a
4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ mean⫾SE ™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 76 5 96⫾10 49⫾7 51 65 5 96⫾12 38⫾8 40 70 77
4 7
117⫾10 107⫾10
36⫾7 61⫾7
31 57
SE⫽standard error.
and vegetable variety (F⫽2.83, Pⱕ.05). The number of items offered varied by school with the two salad bar schools offering four and seven items each, and the preportioned schools offering five items each. Tukey post hoc analysis suggested a difference in consumption between students in the highest and lowest variety schools (P⫽.06), with a small effect size (2⫽0.08). A trend in increasing consumption was noticed as variety increased. Fruit and vegetable weights adjusted for sex and grade levels are presented in Table 2. Salad bars are common in public elementary schools and have become more prevalent in the past 10 years (10). To date no study has reported how the presence of salad bars or fruit and vegetable variety is related to consumption in elementary school students. Our results do not support the hypothesis that salad bars increase fruit and vegetable consumption among elementary school students. We found that students consumed on average 4 g more (⬍3% of a medium apple) fruits and vegetables with a salad bar than students using preportioned servings. These differences are not statistically or practically significant, and are inconsistent with the limited body of salad bar research in schools (17,18). The previous studies were based on either self-report or observational methods. Children’s ability to recall intake is prone to misreporting (19), and observers have been found to overestimate consumption (15,20). Therefore, actual measurements of combined fruit and vegetable weights were used in this study. Measuring the weight of food on each plate produces more precise consumption results than other methods (16,21). We found fruit and vegetable variety of salad bars was associated with greater consumption. Item variety is usually higher with salad bars (10). However, the relation of salad bar variety to consumption has not been evaluated. Schools with the preportioned items offered five items each. Salad bar schools offered four or seven items. The salad bar school with the greatest variety of items offered also had the highest mean amounts of fruits and vegetables consumed, and the school with the lowest variety of items offered had the lowest consumption. These findings suggest that fruit and vegetable variety may mediate the effect of salad bars on fruit and vegetable consumption. Our results provide some suggestions about the mechanisms by which fruit and vegetable variety could increase consumption. Table 2 shows the amount of fruits and vegetables taken was lower in the high-variety salad
bar school compared with the low-variety school, but that the amount consumed at the high-variety school was greater. Perhaps more variety made it more likely that children would find preferred fruits and vegetables. It is surprising that students took larger servings from the low-variety salad bar, but wasted more. This study had several limitations. Because larger differences between serving methods and smaller standard deviations were expected, the power to detect significant differences was limited. A small number of schools were studied, and each school was sampled 1 day only. To control for specific preference of fruits and vegetables and day of the week effects across serving methods, schools were sampled on either a Monday or Friday when they offered similar fruit and vegetable items. Consumption was higher on Mondays compared with Fridays, and higher variety schools were sampled on Mondays and low variety schools on Fridays. Thus, variety and day of the week effects may not be independent. Future studies may want to sample schools on 1 day only. This approach would limit generalizability to other weekdays. An alternative would be to study all schools on all days of the week, which would be preferable and allow for true testing of day-of-week effects. The difference in consumption between salad bars could be due to fruit and vegetable item variety, day of the week, or factors not measured, such as the location of the salad bar, presentation quality, popularity of specific fruits and vegetables, signage, student demographics, or social norms within the school. It should be noted that the school with the highest consumption also had the lowest percent of nonwhite students, so future studies should control for student ethnicity and socioeconomic status. It is possible that present results are not generalizable to other schools. Our results provide initial evidence that fruit and vegetable variety is related to elementary children’s fruit and vegetable consumption. The number of items offered may mediate the effectiveness of salad bars. Future studies should determine whether fruit and vegetable variety affects consumption with both salad bars and preportioned servings while controlling for students’ demographic and economic characteristics. Other potential environmental mediators of salad bar effectiveness, such as location, presentation, and interaction with entrees, also need to be examined.
November 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1791
CONCLUSIONS Environmental interventions combined with educational approaches may be needed to effectively promote children’s fruit and vegetable consumption at school. This study suggests that merely providing a self-service salad bar is not a sufficient intervention to improve elementary school children’s fruit and vegetable intake. However, salad bars with more variety were associated with higher intakes. If confirmed in further studies, salad bars with a high variety of fruits and vegetables may become a recommended strategy for meeting national fruit and vegetable intake goals. This research was funded by a US Department of Agriculture Food Stamp Program award to James F. Sallis, PhD, through the California Nutrition Network (Contract 03-75578). The authors thank the Cajon Valley Union and Lemon Grove School Districts; Linda Patzold, RD, and Patricia McKay, directors of Child Nutrition Services; the cafeteria staff at Magnolia, Ballantyne, Johnson, Mount Vernon, and Vista La Mesa schools; project coordinator Justine Kozo; and research assistants Ayodeji Oladunjoye, Veronica Guzman, Lisa Hopper, and Gary Gailius for their help with this project. References 1. Fruits and Vegetables. Enhanced Federal Efforts to Increase Consumption Could Yield Health Benefits for Americans. Washington, DC: US General Accounting Office; 2002. Publication GAO-02-657. 2. Cummings S, Parham ES, Strain GW. Position of the American Dietetic Association: Weight management. J Am Diet Assoc. 2002;102:1145-1155. 3. US Dept of Agriculture. Nutrition and Your Health: Dietary Guidelines for Americans. 5th ed. Available at: http://www.health.gov/dietaryguidelines. Accessed April 9, 2004. 4. Krebs-Smith SM, Cook A, Subar AF, Cleveland L, Friday J, Kahle LL. Fruit and vegetable intakes of children and adolescents in the United States. Arch Pediatr Adolesc Med. 1996;150:81-86. 5. McGraw SA, Sellers D, Stone E, Resnicow KA, Kuester S, Fridinger F, Wechsler H. Measuring implementation of school programs and policies to promote healthy eating and physical activity among youth. Prev Med. 2000;31(suppl):S86-S97. 6. Weschler H, Devereaux R, Davis M, Collins J. Using the school environment to promote physical activity and healthy eating. Prev Med. 2000;31(suppl):S121S137. 7. Birnhaum AS, Lytle LA, Story M, Perry CL, Murray DM. Are differences in exposure to a multicomponent school-based intervention associated with varying dietary outcomes in adolescents? Health Educ Behav. 2002;29:427-443. 8. Briggs M, Safaii S, Beall DL. Position of the Ameri-
1792
November 2005 Volume 105 Number 11
9.
10.
11.
12.
13.
14.
15.
16. 17.
18.
19. 20. 21.
can Dietetic Association, Society for Nutrition Education, and American School Food Service Association: Nutrition services: An essential component of comprehensive school health programs. J Am Diet Assoc. 2003;103:505-514. Buzby J, Guthrie J. Food Assistance Research Brief— The USDA Fruit and Vegetable Pilot Program Evaluation. Washington, DC: US Dept of Agriculture; 2003. Food Assistance and Nutrition Research Report No. 34-14. US Dept of Agriculture, Food and Nutrition Service, Office of Analysis, Nutrition and Evaluation. School Lunch Salad Bars. Alexandria, VA: US Department of Agriculture; 2003. School Lunch Program: Efforts Needed to Improve Nutrition and Encourage Healthy Eating. Washington, DC: US General Accounting Office; 2003. Publication GAO 03-506. Buzby JC, Guthrie JF. Plate waste in school nutrition programs: Final report to Congress. Economic Research Service E-FAN-02-009. Available at: www. ers.usda.gov/publications/efan02009/. Accessed March 14, 2004. US Dept of Agriculture. The School Nutrition Dietary Assessment Study: School Food Service, Meals Offered and Dietary Intakes. Princeton, NJ: Mathematica Policy Research Inc; 1993. MPR Reference No. 7937-140. Buzzard MI, Faucett CL, Jeffrey RW, McBane L. Monitoring dietary change in a low-fat diet intervention study: Advantages of using 24-hour dietary recalls vs food records. JAMA. 1996;96:574-579. Shankar AV, Gittelsohn J, Stallings R, West KP Jr, Gnywali T, Dhungel C, Dahal B. Comparison of visual estimates of children’s portion sizes under both shared-plate and individual-plate conditions. J Am Diet Assoc. 2001;101:47-52. Kirks BA, Wolff HK. A comparison of methods for plate waste determinations. J Am Diet Assoc. 1985; 85:328-331. Slusser W, Neumann C. Evaluation of the effectiveness of the salad bar program in the Los Angeles Unified School District. Available at: http://socialmarketingnutrition.ucdavis.edu/VIDEO/video2002.htm. Accessed March 14, 2003. Kerfoot B, Fournet R. Evaluation of the use of a self-service fruit and vegetable bar on the consumption and plate waste. J Am Diet Assoc. 1996;96(suppl 9):A-10. Livingstone MB, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc. 2000;59:279-293. Gittelsohn J, Shankar AV, Pokhrel RP, West KP Jr. Accuracy of estimating food intake by observation. J Am Diet Assoc. 1994;94:1273-1277. Comstock EM, Syminton LE. Distributions of serving sizes and plate waste in school lunches. Implications for measurement. J Am Diet Assoc. 1982;81:413-422.