Validity and reproducibility of self-reported intake of fruit and vegetable among 6th graders

Abstract

Objective: Reproducible and valid methods for measuring fruit and vegetable consumption among young school children are urgently needed. The aim of this study was to test the reproducibility and the validity of a newly developed questionnaire on the intake of fruit and vegetable among Norwegian 6th graders. The questionnaire consisted of a 24-h recall part and a food frequency part.

Design: A total of 114 pupils completed the questionnaire two times 14 days apart, and another 85 pupils completed the questionnaire and 7-day food diaries.

Subjects: Pupils of 6th grade with a mean age of 11.9 y.

Results: Spearman correlation coefficients between the frequency part of the questionnaire administered two times varied from 0.62 for fruit to 0.83 for potato, and no difference was seen between the average intakes from the two 24-h recalls. The 24-h recall part of the questionnaire gave higher estimates for the average intake of fruit and juice compared to the 7-day record, while no difference was observed for vegetable intake. Spearman correlation coefficients between the frequency part and the records varied from 0.21 for fruit and potato to 0.32 for the total intake of fruit and vegetable.

Conclusions: Both the 24-h recall and the frequency part gave a consistent response on separate occasions over the test–retest study period. The 6th graders were capable of recording yesterday's intake of vegetable, but overestimated the intake of fruit and juice. The ability to rank subjects based on the frequency part was rather low.

Introduction

A national health priority in many countries including Norway is to increase the consumption of fruit and vegetable to at least five portions per day. Nationwide dietary surveys among Norwegian children and adolescents have shown that the intake of fruit and vegetable is less than half the recommended intake (Johansson & Andersen, 1998; Øverby & Andersen, 2002). Therefore, intervention programmes are needed to increase the intake of these healthy food items. To evaluate the effect of such programmes, reproducible and valid methods for measuring fruit and vegetable consumption are essential.

A variety of different techniques for measurements of habitual dietary intake are available. However, most of these methods are time consuming and demand a certain literacy level and motivation, which makes them unsuitable for young children. In Norway, a school-based fruit and vegetable intervention has been developed (the ‘Fruits and Vegetables Make the Marks Study’). The aim of the study was to increase the intake of fruit and vegetable among 6th graders in primary school. It was desirable to use a dietary assessment method in the intervention study, which could be filled in by the children themselves as part of a school exercise. A questionnaire consisting of a 24-h recall part and a food frequency part was developed. Only a few reproducibility and validation studies have been published evaluating dietary assessment methods for gathering self-reports of intake of fruit and vegetable among children without involving the parents (Domel et al, 1994; Baranowski et al, 1997; Field et al, 1998; Lytle et al, 1998; Edmunds & Ziebland, 2002).

The objective of this paper is to test the reproducibility and the validity of a newly developed questionnaire on the intake of fruit and vegetable among 6th graders. The 7-day precoded food diaries were used as the validity criterion.

Subjects and methods

In total, 129 6th grade pupils from two primary schools in the town of Kongsberg, Norway were invited to participate in the reproducibility study. A total of 125 pupils agreed to participate and filled in the questionnaire during a school hour (Time 1). After 14 days, 116 pupils filled in the questionnaire again (Time 2); two of these had not filled in the questionnaire at Time 1. In the present paper, data from 114 pupils completing the questionnaire at both times are presented (participation rate 88%).

A total of 96 6th grade pupils from three primary schools in Hønefoss, Norway were invited to participate in the validation study, and 87 agreed to participate. The pupils completed the questionnaire during a school hour and were instructed by a nutritionist on how to fill in the precoded food diaries for 7 consecutive days. Two pupils did not complete the study correctly, and data are presented for 85 pupils (participation rate 89%).

In both the reproducibility and the validity studies, project staff administered the questionnaires and responded to questions.

The study protocols were approved by the National Committees for Research Ethics in Norway and the Norwegian Social Science Data Services. Informed consent was sought from both parents and the pupils prior to the study.

Questionnaire

The questionnaire includes a 24-h recall and a short food frequency questionnaire. The 24-h recall part produces data on fruit and vegetable intake on the group level, while the food frequency part is used to rank individuals according to intake. The 24-h recall was completed before the food frequency part. The questions in the 24-h recall part were read out loud one at a time while the participants filled in their responses, for example, the first question ‘Did you eat breakfast yesterday morning?’ was read out loud and the pupils filled in their answers. Then the next question was read out loud and the pupils filled in their answers, etc. This was done in order to help the pupils to recall what they had eaten the day before, and to help them fill in all the questions in this part. Thereafter, the pupils completed the food frequency questionnaire on their own. They were allowed to ask questions. The questionnaire also included questions about other health-related behaviours such as physical activity, smoking, alcohol and weight reduction, and demographic information such as gender and age. In addition, detailed questions on psychosocial determinants were included.

The 24-h recall part aimed to cover total intake of juice, fruit, vegetable and potato for the previous day. French fries were not included in the potato count, but fried potatoes were. The recall part was categorized into five periods: breakfast (before lunch), lunch (at school), after school, dinner and supper (after dinner). For each meal, the participants were asked to record their intake of juice, fruit and vegetable, while the potato intake was only asked for at dinner time. The participants were instructed to give the information in household measures, for example, one apple, two carrots, 10 grapes, glasses of orange juice. Sometimes, the participants used portion to give the information about the amount eaten, for example, a portion of broccoli, a portion of fruit salad. However, the participants did not give any information about the size of the portion. For data analyses, the data were recoded in portions. The conversions from household measures to portion sizes were based on household measures and food weights published by The Norwegian National Association for Nutrition and Health (Blaker & Aarsland, 1989). For example, one apple counts as one portion, one carrot counts as one portion, 12 grapes count as one portion, one portion of broccoli counts as one portion. Since no information about the size of the broccoli portion was available, no distinction could be made according to size. The recoding in the reproducibility and the validation study was carried out by two different persons. To investigate inter- and intravariation according to the recoding, 50 questionnaires were recoded twice by three persons. The intravariation was small: for fruit intake, the correlation between the two times of recoding of the same questionnaires varied from 0.90 to 1.00 for the three persons, while the corresponding figures for vegetable were 0.88–0.94 and 0.98–1.00 for juice. The intervariation was also small: no significant differences were observed between the amount obtained for fruit, vegetable and juice between the three persons recoding the 50 questionnaires (unpublished data).

The food frequency part of the questionnaire included 16 questions about the usual intake of some food items. Seven of these questions were about the intake of the following fruit and vegetable: potato, ‘vegetable for dinner’, ‘vegetable on bread’, ‘other vegetable’, ‘apple, orange, pear and banana’, ‘other fruit and berry’ and ‘fruit juice’. The response categories ranged from ‘never’ to ‘several times a day’, with 10 categories in total. The participants were asked to have their habits during the last 3 months in mind when filling in the food frequency part. The vegetable on bread was not included in the estimation of total vegetable.

Precoded food diary

The participants received seven precoded food diaries and were instructed how to record their entire food intake for 7 consecutive days.

The diary had lists of 277 drinks, food items and dishes grouped together according to a typical Norwegian diet (Øverby & Andersen, 2002). Each food group was supplemented with open-ended alternatives. The design of the booklet was similar to a cross-table, with food listed on the left and time span on the top. Food amounts were presented in predefined household units (eg beverage is recorded in glasses) or as portions estimated from photographs. Along with the diary, each participant had a photographic booklet, which embodied 13 colour photograph series, each with four different portion photographs ranging from small to large. The participants indicate an eating event by filling in how many units of each food item they have eaten in the correct time span. The day is divided into five time spans (from 0600 to 1000, from 1000 to 1400, etc). A copy of the precoded diary can be found at the following website: www.sef.no/assets/11002260/Vedlegg1_ungkost.PDF.

The data entry was made by scanning, using the Teleform program (6.0) (Datascan). The fruit and vegetable data from the diary were recoded into portions. For the food items for which the participants had to use household measures to report their intake, the recoding was carried out in a similar way as for the 24-h recall. For the food items where the participants had to use the photographic booklet, the number of portions eaten was used as the data for portions per day independent of the portion size chosen. This was done to make it comparable with the data from the 24-h recall part where no information about portion size was available.

Statistics

The data were analysed using the computer program SPSS version 11 (SPSS Inc. 1999). Data from the 24-h recall, the food frequency and the 7-day precoded food diaries were generally skewed, hence nonparametric statistical methods were chosen. The sample medians, mean, 25th and 75th percentiles are presented. The differences between the two methods were tested using Wilcoxon's signed-rank test. The ability of the food frequency part to rank individuals was assessed by Spearman's rank correlation coefficient and by classification of subjects into tertiles.

Results

Reproducibility

The mean age of the participants was 11.9 y. There was a higher proportion of girls (n=60) than boys (n=54) in the reproducibility study.

Table 1 shows the daily intake of fruit and vegetable estimated from the 24-recall part at the two time periods. There were no significant differences between intakes estimated at the two periods when looking at fruit, vegetable, juice and potato separately. However, when fruit and vegetable were combined including juice and potato or not, there was a significant difference between the two time periods. This was not observed among boys when analysing boys and girls separately, but among girls fruit and vegetable intake not including juice and potato differed between Time 1 and 2.

Table 1 Intake of fruit and vegetable in portions per day based on the 24-h recall part filled in Time 1 and 2;a n=114, median (P25, P75)b, mean

On comparing the intake estimated from the food frequency part filled in at the two time periods, no significant differences were observed (Table 2). Similar results were observed on analysing girls and boys separately, except for the combined intake of fruit and vegetable among girls. The combined intake of fruit and vegetable was significantly higher the first time that the girls filled in the food frequency part compared to the second time. The Spearman rank correlations ranged from 0.62 for intake of fruit to 0.83 for intake of potato.

Table 2 Intake of fruit and vegetable in times per week based on the food frequency part filled in Time 1 and 2,a and Spearman correlation coefficients with 95% confidence intervals (95% CI); N=114, median (P25, P75)b, mean

Validity

The mean age of the participants was 11.9 y. More girls (n=49) than boys (n=36) were included in the study.

The participants reported a higher intake from the 24-h recall questionnaire compared to the 7-day precoded food diaries (Table 3). The differences in portion per day between the two methods were significant for all items, except for intake of vegetable. The same results were found when the data were analysed for gender separately, except for intake of potato (nonsignificant for boys and girls separately).

Table 3 Intake of fruit and vegetable in portions per day based on the 24-h recall questionnaire and the average 7-day food record; N=85, median (P25, P75),a mean

Figure 1a shows the mean intake of fruit and vegetable (not including potato and juice) from each of the 7 days in the diary recording period. There was a clear decrease in intake from day 1 to the subsequent days. The day-to-day variation in intake varied for the different food items (Figure 1b). In Table 4, the average intake from the 24-h recall is compared with the average intake from the first day in the diary recording period. The average (median) intake from the 24-h recall was significantly higher than the intake from the diary for fruit and juice intake, but not for the intake of vegetable and potato. The differences between the intake from the 24-h recall and intake from the first day of the recording were smaller than the differences observed between the 24-h recall and the average intake from the 7-day diary recording period. However, only the differences for total intake of fruit, vegetable, juice and potato estimated from the 24-h recall and the first day of recording were significantly smaller than the differences estimated from the 24-h recall and the 7-day period averages.

Figure 1
figure1

Day-to-day variation in intake of (a) fruit and vegetable in portions per day (mean), (b) fruit, vegetable, potato and juice separately in portions per day (mean).

Table 4 Intake of fruit and vegetable in portions per day based on the 24-h recall questionnaire and the first day of recording; N=85, median (P25, P75),a mean

The average intake from the food frequency part was significantly higher than that observed from the 7-day food diaries (Table 5). The same results were found when boys and girls were analysed separately, except that there was no significant difference between the two methods with respect to intake of vegetable among girls. The Spearman rank correlation ranged from 0.21 for intake of fruit and potato to 0.32 for total intake of fruit and vegetable. The correlation coefficients were in general higher among girls than boys. The proportion of subjects appearing in the same tertile varied from 35% for intake of fruit to 47% for total intake of fruit, vegetable, juice and potato (Table 6). On average (median), 15% was grossly misclassified.

Table 5 Intake of fruit and vegetable in frequency per day based on the frequency questionnaire and in portions per day from the average 7-day food record, and Spearman's correlation coefficient with 95% confidence intervals (95% CI); N=85, median (P25, P75),a mean
Table 6 Classification of subjects by tertiles of calculated intake from the food diaries and the food frequency questionnaire

Discussion

Measuring food intake is challenging in all age groups, but perhaps especially challenging among children when parents are not involved. In the present study, the reproducibility of a newly developed fruit and vegetable questionnaire used among school children was found to be good. Furthermore, the 24-h recall part of the questionnaire gave valid estimates for the average intake of vegetable, but overestimated the intake of fruit and juice, while the ability to rank subjects according to intake of fruit and vegetable based on the frequency part was rather low.

Reproducibility

The reproducibility of the questionnaire in the present study was good. On the group level, one would expect that the intake from the 24-h recall should be at the same level at Time 1 and 2. No differences were observed between the two time periods for boys and girls combined and boys alone. However, the girls recorded a somewhat higher intake the first time that they filled in the questionnaire compared with the second time. Also, on the frequency part, the girls reported a higher intake the first time compared with the second time, but this was not observed for the boys. The tendency for a higher intake the first time the questionnaire was filled in compared with the second time has been observed in other reproducibility studies (Domel et al, 1994; Buzzard et al, 2001). The correlation coefficients between the frequency parts filled in Time 1 and 2 were high. This was also observed by Eriksen (2001). In the reproducibility studies by Domel et al (1994) and Field et al (1999) among 4–5th and 6–7th grade students, respectively, somewhat lower correlations than those in the present study were observed (0.18–0.54). Correlations of the same size as observed in these two studies were published by Buzzard et al (2001) among 6th graders. It is important to mention that these three reproducibility studies had longer time intervals between administration of the questionnaire compared to our study (Domel et al, 1994; Field et al, 1999; Buzzard et al, 2001).

Validation

In a validation study, the reference method should be as accurate as possible. The energy intake estimated from the diary used as the reference method has been validated among 31 8th graders against energy expenditure estimated from a position and activity monitor. It was found that the energy intake was underestimated with around 20% (Pollestad, 2002). This underestimation of energy intake may also influence the data on fruit and vegetable, and the overestimation observed with the 24-h recall may actually be smaller than that found in the present study. It may appear as if the participants have become tired of filling in the diaries accurately (Figure 1a and b). The intake of fruit and vegetable was highest on day 1 of the diary recording period, and comparing the intake from the 24-h recall with the first day of the record instead of the average from all 7 days results in smaller differences between the 24-h recall part and the records.

In the present validation study, the 24-h recall overestimated the intake of fruit, fruit juice and potato compared to the reference method, while no significant difference was observed for vegetable. There are a few validation studies of the 24-h recall methods used among children, including data on fruit and vegetable (Lytle et al, 1993,1998; Eriksen, 2001). In the two studies by Lytle and co-workers, the 24-h recall data were compared to observational data. They found that among 3rd graders the 24-h recall overestimated the intake of fruit and vegetable; however, this was not observed among 4th graders (Lytle et al, 1993,Lytle et al, 1993, 1998). In a Danish study among 6–10 y olds, it was found that the intake of fruit estimated by 24-h recall was overestimated compared to intake from a 7-day estimated record. Similar to what we found, no differences were observed for vegetable in this study. An important difference between the present study and the Danish study, however, was that in the Danish study, parents helped their children complete the 24-h recalls (Eriksen, 2001).

The overestimation observed with the 24-h recall could, as mentioned earlier, be due to an underestimation of intake by the reference method. Moreover, the pupils may have over-reported the intake in the 24-h recall because they are aware that fruit and vegetable are healthy and socially desirable. The over-reporting was a much larger problem for intake of fruit and juice than for vegetable and potato. An explanation could be that vegetable and potato are usually only consumed as part of one meal in Norway (dinner), while fruit and juice are consumed several times a day; thereby it is easier to over-report the frequency. Furthermore, the pupils may have included the intake of syrup or nectar (fruit juice with some added sugar) when reporting the intake of juice in the questionnaire. It has been observed that children at this age have problems to distinguish these beverages from natural fruit juice.

A number of validation studies of fruit and vegetable questionnaires have been published (for an overview, see van Assema et al, 2002). In most of these studies, the questions on fruit and vegetable were part of a larger questionnaire and only a few of these studies have focused on children (Domel et al, 1994; Baranowski et al, 1997; Field et al, 1998,1999; Eriksen, 2001; van Assema et al, 2002). In the present study, the food frequency part overestimated the intake of fruit and vegetable compared to the diaries. This was also observed by Baranowski et al (1997) and van Assema et al (2002). The correlation coefficients found between the food frequency part and the diaries in our study were low and ranged from 0.21 to 0.32. Table 7 shows that the correlation coefficients obtained for fruit in the present study were similar to what has been observed by other authors. The correlation coefficient for vegetable in our study was lower than that obtained by Eriksen (2001), but similar or higher than those found in the other studies shown in Table 7. This means that even though the correlation coefficients obtained in the present study can be classified as poor, the results obtained are comparable to those obtained in other, similar studies. The number of subjects classified in the correct tertile using the food frequency part varied according to food item. The highest number of correctly classified subjects was observed for the combined intake of fruit, juice, vegetable and potato. The number of pupils misclassified with the food frequency part was relatively high (9–17%). The percentage of correctly classified pupils has not been reported in any of the comparable studies.

Table 7 Correlation coefficients (Pearson or Spearman) observed in validation studies of fruit and vegetable questionnaires among children using records or recalls as the reference method

Questions have been raised whether food frequency questionnaires asking about usual intake is suitable at all for children (Randall, 1991), because they require abstract thinking, as well as basic reading and arithmetic skills, which may be too advanced for young children. Furthermore, children may have difficulties in recalling past events. In a review by McPherson et al (2000) of validation studies conducted among school-aged children, they conclude that recalls and records seem to work better among school-aged children than food frequency questionnaires. The present study confirms the problem using a food frequency questionnaire among young school children for ranking according to intake of fruit and vegetable.

In conclusion, both the 24-h recall and the frequency part of the questionnaire gave consistent response on separate occasions over a 14-day test–retest study period. Moreover, the 6th graders were capable of recording yesterday's intake of vegetable, but overestimated intake of fruit and juice. The ability to rank subjects based on the frequency part was rather low. This has been observed in most studies conducted among children without parental involvement.

References

  1. Baranowski T, Smith M, Baranowski J, Wang DT, Doyle C, Lin LS, Hearn MD & Resnicow K (1997): Low validity of a seven-item fruit and vegetable food frequency questionnaire among third-grade students. J. Am. Diet. Assoc. 97, 66–68.

    CAS  Article  Google Scholar 

  2. Blaker B & Aarsland M (1989): Household Measures and Weight for Food. pp 1–42. Oslo: Landsforeningen for kosthold og helse. (in Norwegian).

    Google Scholar 

  3. Buzzard M, Stanton CA, Figueiredo M, Fries EA, Nicholson R, Hogan CJ & Danish SJ (2001): 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. 101, 1438–1446.

    CAS  Article  Google Scholar 

  4. Domel SB, Baranowski T, Davis H, Leonard SB, Riley P & Baranowski J (1994): Fruit and vegetable food frequency by fourth and fifth grade students: validity and reliability. J. Am. Coll. Nutr. 13, 33–39.

    CAS  Article  Google Scholar 

  5. Edmunds LD & Ziebland S (2002): Development and validation of the Day in the Life Questionnaire (DILQ) as a measure of fruit and vegetable questionnaire for 7–9 year olds. Health Educ. Res. 17, 211–220.

    CAS  Article  Google Scholar 

  6. Eriksen K (2001): Developing and evaluation of a method for measuring changes in intake of fruit and vegetable among children aged 6–10 y. Master thesis, Institute for Human Nutrition. Copenhagen. (in Danish).

  7. Field AE, Colditz GA, Fox MK, Byers T, Serdula M, Bosch RJ & Peterson KE (1998): Comparison of 4 questionnaires assess-ment of fruit and vegetable intake. Am. J. Public Health 88, 1216–1218.

    CAS  Article  Google Scholar 

  8. Field AE, Peterson KE, Gortmaker SL, Cheung L, Rockett H, Fox MK & Colditz GA (1999): Reproducibility and validity of a food frequency questionnaire among fourth to seventh grade inner-city school children: implication of age and day-to-day variation in dietary intake. Public Health Nutr. 2, 293–300.

    CAS  Article  Google Scholar 

  9. Johansson L & Andersen LF (1998): Who eats 5 a day?—Intake of fruits and vegetables among Norwegians in relation to gender and lifestyle. J. Am. Diet. Assoc. 98, 689–691.

    CAS  Article  Google Scholar 

  10. Lytle L, Nichaman MZ, Obarzanek E, Glovky E, Montgomery D, Nicklas T, Zive M & Feldman H (1993): Validation of 24-hour recalls assessed by food records in third-grade children. J. Am. Diet. Assoc. 93, 1431–1436.

    CAS  Article  Google Scholar 

  11. Lytle LA, Murray DM, Perry CL & Eldridge AL (1998): Validating fourth-grade students' self-report of dietary intake: results from the 5 A Day Power Plus program. J. Am. Diet. Assoc. 98, 570–572.

    CAS  Article  Google Scholar 

  12. McPherson RS, Hoelscher DM, Alexander M, Scanlon KS & Serdula MK (2000): Dietary assessment methods among school-aged children: validity and reliability. Prev. Med. 31, S11–S33.

    Article  Google Scholar 

  13. Øverby NC & Andersen LF (2002): UNGKOST 2000. Nation-wide dietary survey among 4th and 8th grade pupils in Norway. Sosial og Helsedirektoratet, Oslo. (in Norwegian).

  14. Pollestad M (2002): Validation of energy intake measured by a pre-coded food diary among 13-y-old students using energy expenditure measured with Actireg. Master thesis, Institute for Nutrition Research, University of Oslo, Oslo. (in Norwegian).

  15. Randall E (1991): Measuring food use in school-aged children. J. School Health 61, 201–203.

    CAS  Article  Google Scholar 

  16. van Assema P, Brug J, Ronda G, Steenhuis I & Oenema A (2002): A short Dutch questionnaire to measure fruit and vegetable intake: relative validity among adults and adolescents. Nutr. Health 16, 85–106.

    Article  Google Scholar 

Download references

Acknowledgements

This project was supported by grants from the Norwegian Research Council. We thank Mona Bjelland, Jorunn Sofie Randby and Marthe Bottolfs for their help with collecting and coding the dietary data.

Author information

Affiliations

Authors

Corresponding author

Correspondence to L F Andersen.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Andersen, L., Bere, E., Kolbjornsen, N. et al. Validity and reproducibility of self-reported intake of fruit and vegetable among 6th graders. Eur J Clin Nutr 58, 771–777 (2004). https://doi.org/10.1038/sj.ejcn.1601875

Download citation

Keywords

  • fruit
  • Norway
  • reproducibility
  • validity
  • vegetable
  • questionnaire

Further reading