Kolodziejczyk, Julia K.*; Merchant, Gina*; Norman, Gregory J.†
Reliable and valid assessment of child and adolescent dietary intake is important for monitoring nutritional status and for conducting epidemiological and clinical research on the relation between diet and chronic diseases (1). Diet quality can be assessed by measuring micro- and macronutrients (eg, vitamins, minerals, fat, fiber) or by measuring consumption of foods and/or food groups (eg, meat, diary, sugar-sweetened beverages). Measuring foods and/or food groups is especially important in instances in which inadequate or overconsumption of certain foods is implicated in disease or ill health such as overweight or obesity (2).
Researchers and clinicians have many options when choosing dietary measures to assess intake of foods and food groups. Standard methods include 24-hour recalls, food diaries, and food frequency questionnaires (FFQs) (1). Often, multiple methods of assessment are used to obtain more precise estimates of dietary intake. Although no dietary measure is without error and limitations (3–5), each method confers different advantages and disadvantages. When choosing from among methods, researchers frequently consider the psychometric properties of the measure(s), but an equally important consideration is the appropriateness of the measure(s) for the study population (6). For example, a recent review found that repeated 24-hour recalls that used parents’ proxy reporting were best for children ages 4 to 11 years for estimating energy intake (7). For children ages 6 to 14 years, weighted food records were superior, and for older adolescents, diet history provided the best estimate of energy intake (7).
FFQs are one of the most widely used dietary assessment tools and are designed to capture usual food intake (1). FFQs are self-report measures that ask participants to report their intake frequency of commonly consumed foods and/or beverages. Participants are presented with a list of foods and are asked to state how often each type of food is consumed (eg, times per day, per week, per month). An estimate of usual or average diet is calculated from participant responses. FFQs are commonly used in research because data analysis is less expensive and time consuming compared with food diaries or multiple 24-hour recalls. FFQs are also less of a burden on participants and have been found to have good psychometric properties (8). FFQs may be preferable to food diaries given that traditional paper food diaries are often falsified (ie, backfilled) and compliance is generally poor (9). Finally, FFQs can be used to measure micro- and macronutrient content as well as intake of foods and/or food groups.
The aim of the present review was to summarize and evaluate the reliability and validity of child and adolescent FFQs that assess food and/or food groups. Although reviews examining dietary intake measurement in children and adolescents have recently been published, these articles focused on dietary assessment of micronutrient intake (10), provided a broad overview of common dietary assessment methods for overweight or obese populations (11), or examined only the validation method of doubly-labeled water, which is most appropriate for energy intake (7). Given the popularity of FFQs, their use in younger populations, and the need to assess intake of specific food and food groups, we saw a need to summarize the reliability and validity evidence to date. This review focuses on food groups from the US Department of Agriculture's “My Plate” dietary recommendations (ie, fruits, vegetables, grains, protein, dairy) (12) as well as other foods commonly consumed by children and adolescents that are often the focus of dietary research (ie, sugar-sweetened beverages and fast foods) (2). Recommendations for best practices when assessing these foods and food groups are proposed based on the review.
A systematic review of child and adolescent FFQ studies was undertaken. To minimize bias and create transparency in our review process, we followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses’ guidelines (13). We searched electronic databases for articles published between January 2001 and December 2010 (last date of search: December 31, 2011). The search was restricted to English-language publications. To conduct a comprehensive search of biomedical and health sciences journals, we used MEDLINE via PubMed (1946–present), the Cochrane Library Online (1999–present), PsycInfo (1806–present), and Google Scholar. Search terms included (food [Title/Abstract] AND frequency [Title/Abstract] AND questionnaire [Title/Abstract]) AND (“Reproducibility of Results”[MeSH] OR “Validation Studies as Topic”[MeSH] OR “Validation Studies”[Publication Type]) AND (“humans”[MeSH Terms] AND Journal Article[ptyp] AND English[lang] AND (“child”[MeSH Terms:noexp] OR “adolescent”[MeSH Terms]) AND (“2001/01/01”[PDAT]: “2011/12/31”[PDAT])). The present review is not registered. Table 1 provides the full list of publication inclusion and exclusion criteria.
Two reviewers independently identified a total of 873 articles. After removing duplicates from across databases, the reviewers separately assessed the eligibility of each article for inclusion. Disagreements were resolved by consensus. To reduce the risk of a study contributing biased validity coefficients to the systematic review, studies that used an FFQ that was administered to the parent, for the child's food consumption, as the validation reference group were excluded. Although we did not undertake additional assessments of the risk of bias within individual studies, readers can surmise the extent to which reliability and validity coefficients were influenced by study characteristics (eg, sample size, population segment, test-retest time interval). For example, the magnitude of a test-retest reliability coefficient will likely be inversely related to the length of the test-retest time interval.
To facilitate the data collection process from our final list of included studies, we developed a data extraction sheet (Appendix 1, http://links.lww.com/MPG/A107). Two reviewers performed the data extraction independently and in a standardized manner. Disagreements were resolved by discussion between the 2 reviewers; if no agreement could be reached, it was planned that the third author would decide. This was not necessary.
Food and Food Groups
The foods and food groups considered in this review include the food groups based on the US Department of Agriculture's “My Plate” dietary recommendations and other commonly consumed foods among children and adolescents (2,15). There were 7 foods/food groups: fruit, vegetables, grains, protein, dairy, sugar-sweetened beverages, and fast foods. Some of these groups had subcategories, which were selected based on the most commonly assessed foods across the studies. The fruit group included 1 subcategory: “100% fruit juice.” The grains group included 9 subcategories: “potatoes,” “bread” (ie, all kinds of bread), “white bread,” “wheat bread,” “cereals,” “pasta,” “rice,” and “whole grains.” The protein group included 8 subcategories: “beans,” “meat” (ie, all meat), “red meat,” “chicken,” “fish,” “eggs,” and “meat substitutes” (eg, a vegetarian food product that approximates the aesthetic qualities and/or chemical characteristics of certain types of meat). The dairy group included 5 subcategories: “milk” (eg, all milk), “skim milk,” “1% milk,” 2% milk,” and “cheese.” The sugar-sweetened beverage group had 2 subcategories: “regular soda” and “diet soda.” Depending on how authors reported the food and food groups, some of the subcategories are reported in clusters (eg, “cereal, rice, and pasta”). In general, the authors’ descriptions of the subcategories were ambiguous and the interpretation of potentially subtle distinctions between “white bread” and “wheat bread” was likely left up to study participants.
Classification and Assessment
The selected studies were grouped according to whether validity or reliability was measured. Validity studies used correlation coefficients to assess the strength of association between the FFQ and a criterion measure (eg, 24-hour recall). Validity studies that assessed a continuous criterion used Pearson product-moment correlation, whereas studies assessing an ordinal criterion used Spearman's rho, and studies that examined agreement between an individual's responses to the FFQ and a criterion measure used kappa.
Reliability studies also used Pearson product-moment correlation, Spearman's rho and kappa; however, in the context of reliability assessment, these correlation coefficients represent the strength of the association between the first administration of the FFQ and the second (ie, test-retest reliability). In some studies, an intraclass correlation coefficient (ICC) was used for reliability. The ICC is the recommended coefficient for test-retest reliability because it assesses accuracy between measures of the same metric (eg, servings of fruits and vegetables), compared with a Pearson coefficient, which assesses consistency of rank order of measures with the same or different metrics (14).
To assess the strength of the relation for both validity and reliability data, the following correlation rating interpretations were used: Pearson and Spearman statistics—0.10 to 0.30 weak, 0.30 to 0.50 moderate, >0.50 strong (15); kappa statistics—<0.40 unacceptable or poor, 0.40 to <0.60 moderate, 0.60 to <0.80 substantial, 0.81 to 1.00 excellent (16); ICC statistics—0.00 to 0.10 virtually none, 0.11 to 0.40 slight, 0.41 to 0.60 fair, 0.61 to 0.80 moderate, and 0.81 to 1.0 substantial (17).
Type of correlation statistic and the correlation coefficient were reported for each food and food group assessed. For validity studies, we reported the correlations between intake as measured by the FFQ and the reference method (ie, for each food, food group, or food subcategory). For each FFQ, we also reported the range of correlations across all of the foods measured, calculated an average correlation, and summarized the magnitude of this average correlation (eg, weak, moderate, strong). For reliability studies, the correlations were reported in a similar manner; they represent only the strength of association between time 1 and time 2 for test-retest reliability. Some studies used >1 type of correlation coefficient to assess reliability and/or validity, which was noted in the summary tables.
The literature review search resulted in 873 articles. After duplicates were removed, 848 articles remained, 821 of which were excluded during the title and abstract screening process. The full text of 27 articles was then examined, 7 of which were excluded for the following reasons: did not report correlations for each food assessed (18–20), the questionnaire assessed eating behavior (21), used an FFQ that was administered to the parent, for the child's food consumption, as the validation reference group (22), and used populations with mean ages below our inclusion criteria (22,23).
After applying the exclusion criteria, 21 articles remained (25–45). The selected studies were classified as to whether validity (N = 18) or reliability (N = 14) was assessed. A total of 11 studies examined both reliability and validity. Ordered by publication year, Table 2 summarizes these articles. It appears that the 21 articles represent 20 different FFQs. The FFQs used by Davis et al and Nelson and Lytle (28,32) were the same; only the questionnaires were tested in different populations. The names of the FFQs are listed in Table 2 (if applicable). Seven publications were from European countries: Belgium (26,36,42), France (29), Austria (34), Germany (34), the Netherlands (43), Sweden (34), Switzerland (34), Denmark (38), Iceland (38), Norway (38), Portugal (38), and Spain (38). Ten articles were from the United States (28,30–33,35,39,40,44,45). Other countries represented in this review include Australia (27), New Zealand (41), Brazil (25), and Uganda (37). A total of 17 publications were validated for adolescents (25,26,28–32,36–45) and 4 were validated for children (27,33–35).
TABLE 2-a Characteri...Image Tools
In 12 publications the FFQs assessed intake of specific types of foods: “beverages” (33), “beverages and fast foods” (28,32), “beverages and snacks” (31), “fruits and vegetables” (25,27,43), “fruit, vegetables, and juice” (30,28,39), “high fat/calories, fruit, vegetables, and grains” (35), and “sugary snacks” (37). The remaining 9 studies focused on “all aspects” of diet (26,29,34,36,40–42,44,45). The size of the FFQs varied from 6–152 items ((36,37), respectively). Seven studies directly assessed portion size (14,15,22,25,27,32,34), 2 included partial portion size information (eg, asked about portion sizes of some items on the questionnaire) (29,32), and 12 did not ask participants about portion size (27,28,30,31,34,35,37,39–42,44).
TABLE 2-b Characteri...Image Tools
The FFQs measured a wide variety of time spans ranging from the previous day to usual intake (ie, no specific time frame was used) and the response options varied from 3 to 10. In terms of method of administration, the majority of FFQs were written (ie, paper based) and administered to the child and/or adolescent (28–32,40,42–45); 5 were filled out by the child or adolescent with help from teachers, parents, or administrators (22,24,26,28); 2 were administered to the parent (27,34); 2 were Web based (15,25); and 1 was administered to the child/adolescent via a personal interview (25). In another study, the questionnaire was written and administered to the child, parents, or caregiver (41).
Eighteen of the 21 studies included validity data (25–40,42,43). These data are summarized in Table 3, ordered by average correlation. Most of the studies used 3- to 7-day diet records (28,30,31,33,36,38,39,42,43) as the validation criterion. The next-most used criterion was either a single (34,35,40) or multiple (25,26,29,32) 24-hour recall(s). Two studies used plasma carotenoids as a biomarker criterion (25,27). The average correlation coefficients across food categories for the FFQs ranged from 0.11 to 0.69. Six studies had weak/poor average correlations (25,28,30,32,34,39), 10 studies had moderate average correlations (26,27,29,33,35–38,42,43), and 2 studies had strong average correlations (31,40).
TABLE 3-a Validity s...Image Tools
The following patterns were found for study characteristics from the validity correlations. Studies that had assessed portion size or partial portion size had validity correlations that ranged from weak/poor to moderate (26,28,31,34,35,38,40,44), but those that did not assess portion size had weak/poor to strong correlations (27,28,30,31,34,35,37,39–42,44). FFQs that assessed food intake within short periods of time (ie, previous day or week) had moderate to strong correlations (31,33,35), whereas those that assessed food intake for longer time periods (ie, 1 month to “usual intake”) had weak to strong correlations (25–30,32,34,36–40,42,43). One study that administered the FFQ via interview had a weak correlation (25). Studies that administered the FFQ to the parents instead of the child (27,34) or that were assisted by a parent, administer, or teacher (33,35,37–39) had weak to moderate correlations. The Web-based FFQs had moderate correlations (15,25) and the correlations for the written FFQs ranged from weak/poor to strong (28–32,40,42–45).
TABLE 3-b Validity s...Image Tools
Fourteen of the 21 studies reported test-retest reliability (26,28,29,31,32,36,38–42,44,45). These data are summarized in Table 4 ordered by average correlation. The time intervals for reliability testing ranged from 7 days to 1 year. The mean reliability correlation coefficients for the FFQs ranged from 0.40 to 0.83. One study had a fair correlation (29), 3 had moderate correlations (37,44,45), and 10 had strong correlations (26,28,31,32,36,38–42). The study with the lowest correlation also had the longest test-retest period of 1 year (29).
Many of the patterns observed in both the FFQ validation and reliability studies are in agreement with previous research. In general, the studies that did not assess portion size (27,28,30,31,34,35,37,39–42,44) had higher correlations with the reference criterion than those that did or those that assessed it partially (26,28,31,34,35,38,40,44). As previous research has suggested, children younger than 8 years may have difficulty estimating portion size because of limited cognitive abilities (47). Although self-report capability is believed to improve to a significant degree by age 12 years (48), the results of this review suggest that assessing portion size may not improve validity because the 2 highest correlations were from questionnaires that did not assess portion size in adolescents. Furthermore, the studies with participants that had a mean age of ≥12 years that did assess portion size had correlations that ranged from weak (25) to moderate (26,36,43).
Similar to other reviews (39), we found that studies with shorter time span assessment periods (ie, assessed food intake from the previous day or week) had better validity (31,33,35) compared with those with longer periods (25–30,32,34,36–40,42,43). Similarly, in the reliability studies, the study with the weakest reliability coefficient had the longest test-retest period of 1 year (29). Earlier work has reported that test-retest correlations are higher when the period is ≤1 month versus ≥6 months (49), and standard practice when conducting test-retest research is to use a period of 2 weeks between each questionnaire administration.
Our review does include some findings that were not reported in previous reviews. First, we observed an association between the number of questionnaire items on the FFQ and the strength of the correlation, which was unexpected because research has suggested that the number of items does not affect validity (50). We found that the 2 studies with the strongest validity correlations had a medium-length questionnaire (ie, 19 items, 63 items) (31,40). These results could indicate that questionnaires must be long enough to assess an adequate amount of information (51) but not inflict too much participant burden because reporting error may result (52).
Second, we did not expect to find an association between method of administration and strength of the validation results because there is limited agreement among researchers as to whether the FFQ should be administered to the parent or the child (53). We found that when the FFQs were administered to the parent instead of to the child, validity correlations were generally lower (27,34) and were only slightly higher when children were assisted by the parent (33). Some suggest that because children younger than 12 years have limited cognitive abilities, it is best for the FFQ to be administered to the parents (47), but others suggest parent administration may not be appropriate because parents are not at school with the child (54).
Several limitations of the present study need to be considered when reviewing the findings. First, owing to practicality, this review could not report all of the validity and reliability correlation coefficients for each type of food assessed in the FFQs. For instance, in the “protein” category, studies reported correlations for specific groups like “hot dogs,” “pork,” “hamburgers,” and so on. As such, readers should refer to the original publication if they seek information on more specific types of food. Second, because only a limited number of studies reported deattenuated or adjusted statistics, we did not report these statistics in this review. In an effort to standardize the correlations and make meaningful and cogent comparisons across studies, only unadjusted correlations were reported. We believe that this is a potential limitation of all of the studies reviewed, given that unadjusted correlations can illuminate, clarify, or nullify existing “known” relations among variables. Third, the reported strengths of the associations from the selected studies are not readily comparable because of the range of different statistics used, and the generalizations that were made should be considered rough estimates.
Fourth, this review did not report comparisons of means between reference methods or percentage of agreement because many of the selected studies did not conduct these types of analyses. Fifth, it should be noted that the definition of food and/or food groups frequently differs across studies, especially among international populations. Interested readers should refer to the original articles to obtain more information about how the authors defined each food and/or food groups. Sixth, this review did not include an assessment of the quality or bias of each study.
FFQs often are used to measure youth dietary intake because they have generally acceptable reliability and validity and are not excessively burdensome for participants. For FFQs to be useful in research or clinical practices, they must be designed for a specific population, not only in terms of culturally relevant food choices but also in design and administration. Researchers or clinicians can use the information in the present review to help determine the most appropriate FFQ to use in their population of interest. In addition, this review sheds light on some important design and administration characteristics to consider regarding the instrument design and administration protocol.
In consideration of the present review's findings and the extant literature, we believe that more research needs to be conducted in this area. The validation studies to date are generally weak. Outcomes and improvements to the assessment of validity could include using more statistical tests (eg, comparison of means). Additionally, the standardization of FFQ studies is warranted so that comparisons across publications can be made more readily. Use of a registration system for validation studies may help in this regard (http://appliedresearch.cancer.gov/cgi-bin/dacv/index.pl). Researchers could also use this registration system to search for FFQs. Given that this review only found 1 FFQ that was validated in a second population (28,32), a registration system that could help researchers find preexisting questionnaires to use in their research would save considerable amounts of time and money. Implementing these recommendations could improve future child and/or adolescent FFQs that would, in turn, advance dietary intake research in this population (Fig. 1).
1. Willett W. Nutritional Epidemiology
, 2nd ed. New York: Oxford University Press; 1998.
2. Barlow SE, Expert Committee. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics
3. Briefel RR, Flegal KM, Winn DM, et al. Assessing the nation's diet: limitations of the food frequency questionnaire. J Am Diet Assoc
4. Sempos C. Invited commentary: some limitations of semiquantitative food frequency questionnaires. Am J Epidemiol
5. Ziegler E, Filer L, eds. Present Knowledge in Nutrition
. Washington, DC: ILSI Press; 1996.
6. Maurer J, Taren DL, Teixeira PJ, et al. The psychosocial and behavioral characteristics related to energy misreporting. Nutr Rev
7. Burrows TL, Martin RJ, Collins CE. A systematic review of the validity of dietary assessment methods in children when compared with the method of doubly labeled water. J Am Diet Assoc
8. Prentice RL, Willett WC, Greenwald P, et al. Nutrition and physical activity and chronic disease prevention: research strategies and recommendations. J Natl Cancer Inst
9. Stone AA, Shiffman S, Schwartz JE, et al. Patient compliance with paper and electronic diaries. Control Clin Trials
10. Ortiz-Andrellucchi A, Henríquez-Sánchez P, Sánchez-Villegas A, et al. Dietary assessment methods for micronutrient intake in infants, children and adolescents: a systematic review. Br J Nutr
2009; 102 (Suppl):S87–117.
11. Collins CE, Watson J, Burrows T. Measuring dietary intake in children and adolescents in the context of overweight and obesity. Int J Obes (Lond)
13. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med
14. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull
15. Cohen J. Statistical Power Analysis for the Behavioral Sciences
. Lawrence Erlbaum; 1988.
16. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics
17. Shrout PE. Measurement reliability and agreement in psychiatry. Stat Methods Med Res
18. Psoter WJ, Gebrian B, Katz RV. Reliability of a sugar consumption questionnaire for rural Haiti. P R Health Sci J
19. Papadopoulou SK, Barboukis V, Dalkiranis A, et al. Validation of a questionnaire assessing food frequency and nutritional intake in Greek adolescents. Int J Food Sci Nutr
20. Prochaska JJ, Sallis JF. Reliability and validity of a fruit and vegetable screening measure for adolescents. J Adolesc Health
21. Penkilo M, George GC, Hoelscher DM. Reproducibility of the School-Based Nutrition Monitoring Questionnaire among fourth-grade students in Texas. J Nutr Educ Behav
22. Roumelioti M, Leotsinidis M. Relative validity of a semiquantitative food frequency questionnaire designed for schoolchildren in western Greece. Nutr J
23. Cade JE, Frear L, Greenwood DC. Assessment of diet in young children with an emphasis on fruit and vegetable intake: using CADET—Child and Diet Evaluation Tool. Public Health Nutr
24. Bennett CA, de Silva-Sanigorski AM, Nichols M, et al. Assessing the intake of obesity-related foods and beverages in young children: comparison of a simple population survey with 24 hr-recall. Int J Behav Nutr Phys Act
25. Slater B, Enes CC, López RVM, et al. Validation of a food frequency questionnaire to assess the consumption of carotenoids, fruits and vegetables among adolescents: the method of triads. Cadernos Saúde Públ
26. Vereecken CA, Bourdeaudhuij ID, Maes L. The HELENA online food frequency questionnaire: reproducibility and comparison with four 24-h recalls in Belgian-Flemish adolescents. Eur J Clin Nutr
27. Burrows TL, Warren JM, Colyvas K, et al. Validation of overweight children's fruit and vegetable intake using plasma carotenoids. Obesity (Silver Spring)
28. Davis JN, Nelson MC, Ventura EE, et al. A brief dietary screener: appropriate for overweight Latino adolescents? J Am Diet Assoc
29. Deschamps V, de Lauzon-Guillain B, Lafay L, et al. Reproducibility and relative validity of a food-frequency questionnaire among French adults and adolescents. Eur J Clin Nutr
30. Di Noia J, Contento IR. Use of a brief food frequency questionnaire for estimating daily number of servings of fruits and vegetables in a minority adolescent population. J Am Diet Assoc
31. Neuhouser ML, Lilley S, Lund A, et al. Development and validation of a beverage and snack questionnaire for use in evaluation of school nutrition policies. J Am Diet Assoc
32. Nelson MC, Lytle LA. Development and evaluation of a brief screener to estimate fast-food and beverage consumption among adolescents. J Am Diet Assoc
33. Marshall TA, Eichenberger Gilmore JM, Broffitt B, et al. Relative validity of the Iowa Fluoride Study targeted nutrient semi-quantitative questionnaire and the Block Kids’ Food Questionnaire for estimating beverage, calcium, and vitamin D intakes by children. J Am Diet Assoc
34. Michels KB, Waser M, Ary E, et al. Validation of a questionnaire to assess dietary habits among 5-13-year old school children of farmers and anthroposophic families. J Nutr Environ Med
35. Thiagarajah K, Fly AD, Hoelscher DM, et al. Validating the food behavior questions from the elementary school SPAN questionnaire. J Nutr Educ Behav
36. Matthys C, Pynaert I, Keyzer WD, et al. Validity and reproducibility of an adolescent web-based food frequency questionnaire. J Am Diet Assoc
37. Kiwanuka SN, Astrøm AN, Trovik TA. Sugar snack consumption in Ugandan schoolchildren: validity and reliability of a food frequency questionnaire. Community Dent Oral Epidemiol
38. Haraldsdottir J, Thorsdottir I, de Almeida MD, et al. Validity and reproducibility of a precoded questionnaire to assess fruit and vegetable intake in European 11- to 12-year-old schoolchildren. Ann Nutr Metab
39. Cullen KW, Zakeri I. The youth/adolescent questionnaire has low validity and modest reliability among low-income African-American and Hispanic seventh- and eighth-grade youth. J Am Diet Assoc
40. Hoelscher DM, Day RS, Kelder SH, et al. Reproducibility and validity of the secondary level school-based nutrition monitoring student questionnaire. J Am Diet Assoc
41. Metcalf PA, Scragg RK, Sharpe S, et al. Short-term repeatability of a food frequency questionnaire in New Zealand children aged 1-14 y. Eur J Clin Nutr
42. Vereecken CA, Maes L. A Belgian study on the reliability and relative validity of the health behaviour in school-aged children food-frequency questionnaire. Public Health Nutr
43. Van Assema P, Brug J, Ronda G, et al. A short Dutch questionnaire to measure fruit and vegetable intake: relative validity among adults and adolescents. Nutr Health
44. Buzzard IM, Stanton CA, Figueiredo M, et al. Development and reproducibility of a brief food frequency questionnaire for assessing the fat, fiber, and fruit and vegetable intakes of rural adolescents. J Am Diet Assoc
45. Speck BJ, Bradley CB, Harrell JS, et al. A food frequency questionnaire for youth: psychometric analysis and summary of eating habits in adolescents. J Adolesc Health
46. Deleted in proof.
47. Livingstone MB, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc
48. Livingstone MBE, Robson PJ, Wallace JMW. Issues in dietary intake assessment of children and adolescents. Br J Nutr
2004; 92 (suppl 2):S213–S222.
49. Cade J, Thompson R, Burley V, et al. Development, validation and utilisation of food-frequency questionnaires-a review. Public Health Nutr
50. Cade JE, Burley VJ, Warm DL, et al. Food-frequency questionnaires: a review of their design, validation and utilisation. Nutr Res Rev
51. Serdula M, Byers T, Coates R, et al. Assessing consumption of high-fat foods: the effect of grouping foods into single questions. Epidemiology
52. Livingstone M, Prentice A, Coward W, et al. Validation of estimates of energy intake by weighed dietary record and diet history in children and adolescents. Am J Clin Nutr
53. Collins CE, Watson J, Burrows T. Measuring dietary intake in children and adolescents in the context of overweight and obesity. Int J Obes (Lond)
54. Baranowski T, Sprague D, Baranowski JH, et al. Accuracy of maternal dietary recall for preschool children. J Am Diet Assoc